Development of carbon monoxide releasing molecules for the treatment of rheumatoid arthritis

Final Report Summary - CORMRA (Development of carbon monoxide releasing molecules for the treatment of rheumatoid arthritis)

Executive Summary:
The CORMRA project focused on the pre-clinical development of carbon monoxide releasing molecules (CORMs) for the treatment of rheumatoid arthritis (RA). Carbon monoxide (CO) has been known to exert a very powerful anti-inflammatory effect and specifically acts upon a number of cellular and molecular mechanisms that are involved in the RA disease process. Although there is currently no cure for RA, several types of drugs are prescribed for this disease in Europe and elsewhere. These include steroids, non-steroidal anti-inflammatory drugs (NSAIDs) and disease-modifying anti-rheumatic drugs (DMARDs). All these drugs present severe side effects that limit their broad use and are very far away from representing the optimal therapy to RA. There is a great need for effective, harmless and affordable drugs that treat RA. CORMs hold the promise to become exactly that.
The administration of CO as a gas to RA patients would be impractical and not effective, especially in light of the affinity with which CO binds to haemoglobin and of the difficulties involved in repeated administration of a gaseous treatment to patients with a chronic disease.
CO-releasing molecules (CORMs) promise to harvest the therapeutic potential of CO by delivering it in a more practical and controlled fashion. The classes of CORMs used in this project are small molecules, some of which release CO only or preferentially in chemical environments that are typical of an inflamed site, such as arthritic joints. In addition, some CORMs can be administered orally, and most of them can be chemically modified to have better pharmacological properties without losing their CO releasing ability. Some CORMs had already been tested in preliminary experiments using rodent models of RA, and had yielded encouraging results in terms of efficacy and toxicity.
The main goals of this project were:
• to select a small number of exceptionally promising CORMs for the treatment of RA
• to test their detailed pharmacological and toxicological properties
• to arrive at one or more lead candidates for clinical development
With these aims, three different SMEs joined together their resources, critical mass and knowledge in the field of CORM chemistry, pharmacological testing, in vitro and in vivo disease models, and standardization and validation of assays.

The work was divided between the partners as follows:
• Chemical synthesis of CORMs (in gram-scale) to be tested; synthesis of structurally optimised molecules on the basis of SARs (Structure-Activity Relationships) derived from the biological tests (Alfama)
• Chemical testing (stability, CO release kinetics, solubility) (Alfama, Frimorfo, SBQ, Micron)
• In vitro toxicity and preliminary efficacy (Alfama, UNICT)
• In vivo acute toxicity (UNICT, Alfama)
• In vivo inflammation and rheumatoid arthritis models (UNICT, IMM, FRIMORFO)
• Ex vivo anti-inflammatory effects on human samples collected from rheumatoid arthritis patients (IMM)
• Histological, biochemical and haematological analyses (Frimorfo, SBQ)
• Validation of efficacy, metabolism, distribution, pharmacokinetics, bioavailability (Frimorfo, SBQ, Micron)
• Formal toxicology, GMP manufacturing, Submission of IND application and preparation and performance of clinical trials (all post-project) (Alfama)

As a consequence of this work, a lead compound was selected as promising anti-arthritic CORM: a molecule called ALF826. Ex vivo this molecule seems to affect the activation and function of neutrophils and macrophages isolated from blood samples collected from rheumatoid arthritis patients.

Project Context and Objectives:
Rheumatoid arthritis (RA) is a highly incapacitating chronic inflammatory disease of the joints, which affects roughly 1% of adults in developed countries and currently has no cure. Several types of drugs are prescribed for this disease in Europe and elsewhere, including steroids, non-steroidal anti-inflammatory drugs (NSAIDs) and disease-modifying anti-rheumatic drugs (DMARDs). All these drugs present severe side effects that limit their broad use and are very far away from representing the optimal therapy to RA. There is a great need for effective, harmless and affordable drugs that treat RA. Because carbon monoxide (CO) has been shown to have strong anti-inflammatory effects if given at low doses for continuous periods of time, molecules that can deliver CO to the joints in a safe and controlled manner will hold the promise to treat this grave disease.
The final objective of the project was to prepare new CO-Releasing Molecules (CORMs) that show efficacy in in vitro and in vivo models of rheumatoid arthritis (RA) and inflammation, and to be able to select at least one such molecule for subsequent clinical development.
The overall strategy of the work plan was to generate a large variety of molecules with a potential to release CO, and select the best ones through a number of tests performed, in a sequential manner. The compounds were progressively selected for their pharmacologic and therapeutic properties, namely their efficacy in treating inflammation and arthritis in in vitro and in vivo models without causing undesired side effects, and displaying an adequate pharmacological profile. An iterative synthesis-test-synthesis loop was used to optimize the pharmacological and therapeutic profiles.
During the first year the main objectives of the project consisted in producing a relatively large number of compounds and testing their basic chemical and pharmacological properties (in vitro), in order to be able to choose a first batch of “suitable” compounds for testing in animal models of arthritis. The results obtained in the first year were very encouraging, in view of the fact that the first in vivo tests (in an Adjuvant Induced Arthritis model) clearly eliminated two classes of compounds that had low acute toxicity and excellent stability profiles: those with phosphine and isocyanide ligands. The latter actually proved to excel in another disease model, of acute liver failure (outside this project), but were ineffective in RA model.
Therefore, the work of the last eight months into the second year of the project was focused only on the other two classes of compounds – with diimine ligands of the PyCa and DAB types. The latter did not enter in vivo tests for lack of time and in view of the encouraging results of the PyCa type complexes. Those results led five compounds into the much more advanced K/BxN arthritis model, producing very encouraging results for two of them, at which time the activity of the project was suspended.

Project Results:
The main project results were:
1) Chemical synthesis, evaluation of stability, CO release kinetics, solubility, biological evaluation and selection of the best candidate CORMs to treat RA
2) Validation and standardisation of several biological and chemical assays.

Chemical synthesis, evaluation of stability, CO release kinetics, solubility biological evaluation and selection of the best candidate CORMs to treat RA
The work that led to these results was focused on the preparation of CORMs and on the testing of their basic chemical and biological properties. This allowed a selection of a small number of CORMs for testing in in vitro and in vivo models of inflammation and of rheumatoid arthritis.
This work covered different families of compounds with different ligand settings and CO release profiles. A selection of these CORMs was screened through in vitro chemical and cell tests and in vivo models of inflammation and rheumatoid arthritis. The results obtained at the end of the first 12 months eliminated several families of compounds and led us to concentrate on the development of two families of metal-based CORMs with low toxicity and encouraging in vivo tests.
Thus, a second wave of synthetic work designed to improve and optimise the anti-arthritic profile of such complexes through the variation of the ligand functionalization was carried out to the end of the first 20 months.
New in vitro and in vivo studies performed at IMM with these more advanced CORMs suggested ALF826 as a promising anti-arthritic CORM molecule and likely the best candidate for further studies. This molecule was also seen to affect, ex vivo, the activation and function of neutrophils and macrophages isolated from blood samples collected from rheumatoid arthritis patients.

Validation and standardisation of several biological and chemical assays.
During the project, some in vitro assays were developed by different RDT performers. Micron RS had the responsibility to standardise and validate some of these assays. Micron RS selected two inflammation models set up by elements from Participant 4 (Dr D. Boraschi) to be standardised. In the first year of the project Standard Operating Procedures (SOP) were issued for the Buffy Coat preparation, PBMC preparation, monocyte isolation and evaluation of monocyte concentration. In the second year Micron RS issued, in collaboration with another P4 element (CNR), other Standard Operating Procedures regarding the quantification of human IL-1 and IL-6 from culture supernatants, the isolation of total RNA from cultured cells and the quantification of panels of human cytokines and chemokines in culture supernatants.
Several experiments were also conducted in collaboration to standardise an assay developed by Participant 1, Alfama. This assay allows determining the activity and cytotoxicity of CO releasing molecules (CORM). A clear reduction of the cell proliferation was observed, but it was not very reproducible in different experiments. The standardization protocols represent the deliverables for WP7.

Detailed Description of Objectives and Results
More specifically, in accordance with Annex I to the Grant Agreement, the objectives of the project were the following:
a) Synthesis of improved CORMs derived from the suggestions from the biological tests. Studies on stability, solubility and kinetics of CO release of the synthesised CORMs (WP1).
b) Evaluation of toxicity and efficacy of CORMs in vitro and in vivo, evaluation of the effects of the compounds on the M1 vs. M2 activation pathways of human monocytes and mouse resident peritoneal macrophages, testing of general anti-inflammatory properties, measurement of COHb levels after administration of compounds in vivo (WP2)
c) Pharmacokinetics of different CORMs and/or CORMs formulations in the rat adjuvant induced arthritis models (WP3)
d) Proof of concept studies of lead and optimised CORM in preclinical models of rheumatoid arthritis (RA) (WP4)
e) Evaluation of the anti-inflammatory effect of the administration of a CORM in a mice model of arthritis (WP5)
f) Validation and Standardisation of Analytical Methods (WP7)
g) Dissemination and Training (WP8)
h) Intellectual Property Protection and Project Management (WP9)

a) Synthesis of CORMs to be tested. Studies on stability, solubility and kinetics of CO release of the synthesised CORMs (WP1).
As a consequence of efficacy, stability and toxicology results obtained after the project proposal was submitted but before the work began, it was decided that only organometallic compounds would be included in the project, and thus organic compounds (tertiary aldehydes), which had been mentioned in the application, were excluded from the initial selection.
During the first year four major families of complexes were developed, with different types of L ligands. These included amines (NR1R2R3), phosphines (PR1R2R3), isocyanides (CNR) and diimines. The latter comprise 2 sub-classes, one derived from pyridine 2-carboxyaldehyde (PyCa; 2-imino-carboxy-pyridine; Py-2-CR=N-R’) and another derived from 1,3-diazabutadiene (DAB; R’-N=CR-RC=N-R”). The compounds were completely characterised by Elemental Analysis (C, H, N); NMR spectroscopy (1H-NMR and 13C-NMR when necessary); IR spectroscopy. This first effort led to the initial preparation of ca. 50 compounds in the first 3 months of the project.
The stability of these compounds in biologically compatible media was evaluated by HPLC and by the measurement of CO release, whereas the solubility was measured through standard techniques. Several HPLC protocols were established that enable checking the presence of impurities and the decay of the CORM in solution along the time. This decay may or may not imply release of CO. Insoluble CORMs were formulated as emulsions using Cremophore EL and methylpyrrolidone in saline for in vivo testing.
The rate of CO release was evaluated in vitro: compounds were incubated in solution in closed vials and CO determined by gas chromatography (GC). This general method encompassed two variations: one using GC-TCD detection and another using the highly sensitive GC-RCP (Reduction Compound Performer). The former methodology was used for the least stable complexes whereas the latter was used for those that barely decompose in aqueous media. In parallel another methodology was introduced: measuring the CO release to whole blood in vitro through standard oximetry. This method gives an idea of the stability of the CORMs in blood. This is crucial information for the selection process because it allows the identification and subsequent elimination of those compounds that rapidly decompose in blood and produce high values of COHb (carboxyhemoglobin) in circulation, which are not acceptable clinically. Besides stability, the control of water solubility in these complexes was also a major concern given the high lipophilicity of the Mo(CO)x (X = 3, 4, 5) fragment.
These types of tests rapidly led to the exclusion of the complexes bearing amines as ancillary ligands (8 compounds) since they showed very low stability in aqueous systems. The other families of compounds were found to have sufficient stabilities in aerobic, biocompatible solutions and allowed the selection of 20 compounds to pass on to cytotoxicity studies and biological assays in vitro in WP2.
This selection encompassed all four different types of ligands: isocyanide, phosphine, PyCa and DAB as well as different values of CO according to the general formula Mo(CO)nL6-n (n = 3, 4 ,5). The ligands clearly modulate stability, solubility and CO release profiles in vitro. With the exception of the phosphines, the other classes of ligands provided large room to vary the “decoration” of the ligands. Such variations affect not only the CO releasing profile of the complexes but also their ADME (Administration, Distribution, Metabolism and Excretion) characteristics and pharmacological efficiency.
This task, within WP1, was finished in the first year and all the deliverables were produced.
As the results of the biological tests in WP2, a shorter list of five compounds (ALF496, ALF541, ALF575, ALF586, ALF587) was tested in the in vivo model of rheumatoid arthritis (AIA model; see WP4). The results excluded the phosphine and isocyanide complexes and pointed to the good performance of the PyCa and DAB complexes. This led to a second wave of synthesis of complexes bearing this type of ligands, which lasted to the end of the first year of activity and even beyond. A total of 125 complexes prepared within this task (see Table WP1.1).

b) Evaluation of the toxicity and efficacy of CORMs in vitro and in vivo; evaluation of the effects of the compounds on the M1 vs M2 activation pathways of human monocytes and mouse resident peritoneal macrophages; testing of general anti-inflammatory properties; measurement of COHb levels after administration of compounds in vivo (WP2).

General toxicity and anti-inflammatory properties in vitro.
During the first year of the project, the 20 compounds selected on the basis of stability, solubility and CO release profile in vitro were further screened by examining cell toxicity and activity in an in vitro model of inflammation. From all the compounds analysed the most toxic were the isocyanide complexes. However, that toxicity was not a function of the number of CO ligands in the molecule, and seemed to depend on the actual functionalisation of the ligand. This kind of variation is also present within the phosphine complexes. However, the complexes with PyCa or DAB ligands were found to be almost non-toxic to RAW264.7 cells (survival between 90-100%). This low toxicity is kept almost without exception for compounds of the PyCa ligand family (30 examples; Table WP1.1)
The anti-inflammatory properties of the compounds were tested in vitro using two tests that have been classically used in all the published research on CORMs: the inhibition of NO and TNF-α from RAW264.7- stimulated with LPS (Lipopolysaccharide).
The complexes with PyCa-ligands were the least effective complexes tested in this in vitro model. These compounds were not able to significantly decrease the levels of NO or TNF-α produced by LPS-activated RAW264.7 cells. This observation contrasts with their encouraging activity observed in vivo and in other in vitro studies (see two next subsections and WP4 and WP5).
The two DAB-ligand complexes were able to decrease NO levels to approximately 50% of the control levels but were not able to decrease TNF produced by RAW264.7 induced cells.
The phosphine-ligand complexes were also not active in reducing the NO produced by RAW264.7- induced cells. However, ALF571 and ALF760 were able to decrease TNF levels in activated macrophages in a dose-dependent manner. In particular, ALF571 was able to decrease TNF levels by 40%. These results contrast with the toxicity of these compounds observed in vivo (see WP4)
The isocyanide-ligand complexes were able to reduce both NO and TNF levels induced by LPS in a dose-dependent manner. From the complexes of this family the most effective in reducing the levels of the two pro-inflammatory mediators was ALF586, which was able to reduce the NO production by 87% and the TNF production by 56%. The strong inhibition obtained with ALF589 may be due to cell death, because the compound is rather cytotoxic.
The effects of CORMs on red blood cells (RBC) must be kept in mind during development stages because RBC haemolysis is highly deleterious. Therefore, all compounds were tested according to a standard literature procedure for the determination of the haemolytic index. This assay was performed using sheep red blood cells (SRBC) obtained after centrifugation of sheep whole blood. From the compounds tested on the haemolysis assay only the DAB group complexes, ALF555 and ALF758, were toxic to sheep red blood cells in concentrations below 1 mg/mL. All the other compounds tested in this assay were not haemolytic up to 1 mg/mL concentration.
The combination of these assays led to a first selection of 10 compounds which were proven to be non-toxic (at reasonable concentrations) and inhibit pro-inflammatory cytokine production in vitro, providing insights into mechanistic mode of action of CORMS on inflammatory pathways of macrophage activation. This selection comprised 4 phosphine derivatives (ALF496, ALF541, ALF571, ALF760), three isocyanide derivatives (ALF586, ALF587, ALF749), 2 PyCa derivatives (ALF574, ALF575) and one DAB derivative (ALF555). A further analysis of this selection taking into account solubility and other pharmacological parameters narrowed the choice to five compounds for the first exploratory tests of the treatment of RA, using a rat model of AIA (see WP4). These compounds, ALF496, ALF541, ALF586, ALF587 and ALF575, cover all groups except the DAB scaffold which was removed due to its slightly haemolytic properties, which could pose problems in a rather time prolonged test. As already mentioned above, the performance of these five compounds in the AIA model of RA, led to the selection of PyCa derivatives for the next rounds of in vivo RA tests (see WP4 and WP5).

Effect on monocytes – selection, set-up and adaptation of in vitro model protocols
The study on monocytes encompassed a period of about six months for the set up and validation of the assays, before the beginning of the assessment of the activities of the CO-releasing compounds. The assays have a double scope: a) assess the anti-inflammatory activity of the CORMs on human inflammatory cells in systems that allow comparison between physiological and pathological inflammation; b) investigate the mechanism of action of the best compounds in inhibiting the inflammatory activation/activity.
In the first four months of activities (July 1st, 2010 – October 31st, 2010), one of the two in vitro models was studied and implemented, i.e. the kinetic stimulation of human primary monocytes in vitro in conditions mimicking the initiation, development and resolution of the physiological defence inflammatory response. The second model, recapitulating the onset, development and chronicity of pathological inflammation in Rheumatoid Arthritis (RA) was expected to provide information on the capacity of CORMs to inhibit pathological inflammation vs. the physiological defence response (in search of molecules able to inhibit chronic inflammation while allowing the physiological defence response to take place), and on their mechanisms of action (by complete response profiling for the most interesting molecules). The fact that the model systems are based on human primary cells, as opposed to animal cells or continuous cell lines, adds reliability and representativeness to the results.

Effect on human monocytes and mouse resident peritoneal macrophages
The human monocyte inflammation model encompasses the kinetic evaluation of activation parameters in the phases of initiation of inflammation, development, and resolution. The model has the advantage of being based on human primary cells (as compared to in vivo models in induced inflammation in mice or rats and to in vitro assessment with mouse macrophage-like tumour cells like the RAW264.7) thus avoiding the substantial differences observed in the inflammatory responses of humans vs. rodents, and validating data obtained in the other systems. In the second period, the anti-inflammatory effects of CO-releasing compounds provided by Alfama was examined in two experimental ex vivo settings, one reproducing the development of physiological inflammation, and a second one mimicking a persistent inflammation. The third model, which reproduces a pathological RA-like inflammation, has not been used yet.
Human CD14+ blood monocytes were isolated from PBMC obtained from buffy coats of healthy donors upon gradient separation. The detailed standardised protocol of cell isolation and characterisation has been prepared together with Dr. Claudio D’Urso (Micron). Replicate cultures were set up with 5x106 monocytes/well in 2 ml of RPMI-1640 medium supplemented with 5% human AB serum. Monocytes were exposed kinetically to a sequence of conditions simulating the initiation, development and resolution of the inflammatory response. At time zero monocytes were exposed to CCL2 (20 ng/ml) that simulates recruitment to the site of inflammation and incubated at 37°C for 2 h. CCL2 was then washed off, and cells were exposed to LPS (5 ng/ml) and incubated at 39°C for one additional hour, then TNFα (10 ng/ml) was added and incubation continued for 4 h. At time +7 h, IFN-© (25 ng/ml) was added and incubation at 39°C continued for additional 7 h. At time +14 h, the inflammatory stimuli were washed off, fresh medium added, and cells were either returned to 37°C (physiological resolution of inflammation) or left at 39°C (no resolution, persistent inflammatory reaction). Supernatants and cells were collected for analysis at time zero (freshly isolated monocytes, only cells), + 4 h (early inflammation), + 14 h (late inflammation) and + 24 h (resolution or persistent inflammation). Additional samples were collected at 7 and 9 h in the model of persistent inflammation. Analytical endpoints were gene expression (real-time PCR) and protein production (ELISA) for IL-6, IL-1®, IL-1Ra, IL-18 and IL-18BP.
Two treatment schedules were designed: A. Pretreatment. Drugs were added to the culture medium at time zero (2 h before the early inflammatory stimulation), time +5 h (2 h before the late inflammatory stimulation), and time + 14 h (at the beginning of the last phase). B. Concomitant treatment. Drugs were added to the culture medium at times +2 h (together with the early inflammatory stimulation), +7 h (together with the late inflammatory stimulation) and + 14 h (at the beginning of the last phase).
Drugs were solubilised and used according to the protocols designed by Alfama. Preliminary trials showed that the volumes of DMSO, HEPES buffer and sodium bicarbonate added to the culture did not significantly affect the cell viability and function. Dexamethasone (DXM; 10 µM) was used as control.
The preliminary results of cytokine production show that the CO-releasing compounds ALF554, ALF826 and ALF845 display clear activity, which can be observed not only when the compounds are administered before or at the time of inflammatory stimulation, but also when given at later times (5 h) when the inflammatory reaction is already on-going. It should be noted that, as compared to DXM, the CO-releasing compounds are less effective in inhibiting the anti-inflammatory cytokine IL-1Ra, thus adding to their anti-inflammatory capacity. Further studies will be important. Detailed information on these results is reported in WP2, and confirmation of the potential of ALF826 in WP5.

Measurement of COHb levels after administration of compounds in vivo
The complexes were also evaluated in terms of their CO release profile in vivo. The values of COHb they produce in circulation, after i.v. administration, are not very elevated and most of them fall below the 10% COHb threshold at 10 min post administration, which is when the peak is usually reached. The measurements are carried out by taking samples of blood from the animals and measuring them in a conventional oximeter. As expected, the values of COHb in circulation depend on the nature of the ligand. Phosphine and isocyanide derived compounds always produce values < 8 %COHb in systemic circulation. These values are higher than those found in the in vitro experiments with whole blood, which may be due to metabolic modification of the complexes in vivo, which in turn increases the rate and extension of CO release. In fact, in the course of another project at Alfama this was shown to be the case for isocyanide complexes. For ethical reasons, such tests were only performed after having selected a given compound for in vivo testing in order to make sure that it didn’t generate excessive CO spill over to the systemic circulation.
Very interestingly, complexes ALF821 and ALF826, which preformed very well in the K/BxN model used in the most advanced stages of this project, produce rather low CO in circulation.

c) Pharmacokinetics of different CORMs and/or CORM formulations in mouse collagen and/or the zymosan induced arthritis models (WP3)

Testing of selected CORMs/formulations for the plasma response in mice
Participant SBQ investigated the chromatographic and mass spectrometric behaviour of two different CORMs. HPLC-MS methods were developed for these CORMs for their quantitation in mouse plasma and serum. The mass spectrometric sensitivity and stability of CORMs was investigated in solution as well as in mouse plasma. It was observed that sensitivity of the analytical method is compound-dependent but in all investigated cases a limit of detection between 1.00 and 10.0 ng/mL could be achieved. A method was developed for the quantitation of ALF795 in mouse serum and the concentration of ALF795 was determined in samples from an acetaminophen (APAP) induced acute liver failure test. This method constitutes an approach to methods to be developed for other compounds.
In addition, microsomal incubations were performed using rat liver microsomes over a time period of 60 minutes. The microsomal stability of four CORMs was investigated and the decrease of the incubated compound was determined by HPLC-MS using electrospray ionisation. In addition, phase I microsomal metabolites were studied but only in one CORM an extensive metabolism and ligand related metabolites were found and identified.
Three additional CORMs, ALF521, ALF528 and ALF545, were investigated by incubation with rat liver microsomes. ALF528 and ALF545 were degraded without the addition of microsomes whereas ALF521 was stable in the absence of microsomes. In the presence of microsomes, approximately 50% of ALF521 was metabolised after an incubation time of 10 minutes. No metabolites were identified for the CORMs analysed.

Test of selected compounds/formulations for bioavailability at the target organ (joint), and for efficacy
Unfortunately, due to the premature termination of the Project no CORMS could be tested in the second line test and the joint histology.

Establishment of the technology for measurement of the CO release by the CORM by semi-quantitative immunohistochemistry of the joint for CO-heme proteins (e.g. CO-cytochrome C oxidase)
Due to the premature termination of the project, this task could not be achieved.

d) Proof of concept studies of lead and optimized CORM in preclinical models of rheumatoid arthritis (RA) (WP4)

Testing of selected CORMs in rodent models of arthritis
In the first year of the project, five compounds from three classes of molybdenum carbonyl complexes were evaluated by the group of Prof. Ferdinando Nicoletti at the Università di Catania (see below), in in vitro models of pathological inflammation and in the rat model of adjuvant-induced arthritis (AIA).
As a first approach to the testing of selected CORMs in an animal model of arthritis, five molecules arising from the previous tasks (ALF541, ALF496, ALF587, ALF586 and ALF575; Table WP2.3) were assessed for their in vivo toxicity and tested in a rat model of adjuvant induced arthritis (AIA) at the University of Catania. The tests were performed through the administration of several doses of each selected compound to determine lethality and/or other visible, immediate side effects on the animals’ behaviour, level of activity and general health. Animal sacrifices necessary at this stage will avoid unnecessary deaths at later stages. ALF541 and ALF496 proved to be toxic. ALF587 was less toxic. The other two compounds, ALF575 and ALF586 were well tolerated. ALF575, a PyCa based compound, was associated with a significant reduction of both paw oedema and arthritic scores. Although the precise mode of action of ALF575 in this model has not presently been ascertained, it is of interest that during the one-month follow-up after treatment interruption none of the rats that had been treated with ALF575 developed clinical signs of AIA. This is at striking contrast with what we have observed in the positive control group of rats that had been treated with Dexamethasone, where we observed a late occurrence of AIA in 4 out of 5 during the follow-up period. Although speculative at this stage, this observation may suggest that Dexamethasone and ALF575 possess different immunopharmacological modes of action that may eventually render them suitable for additive or synergistic effects.
Following these tests in the first year of the project, which dictated the elimination of phosphine and isocyanide based ligands, a wealth of Mo(CO)4PyCa type compounds, was pushed forward to the front of the screening process. So, in the second period, ten molecules, all belonging to the PyCa type of complexes, were tested at the University of Catania, to evaluate the possible anti-inflammatory effects in the rat model of AIA. The treatments with all the tested compounds for 30 consecutive days were well tolerated by the rats as judged by their clinical status and body weight increase compared to vehicle treated rats. The rats treated with ALF821 exhibited a clinical course of the disease significantly lower than the vehicle treated rats and this effect seemed to be long-lasting and not dependent on the prolonged application of the compound as the rats were still protected from clinical development of AIA 16 days after treatment interruption.
Also the rats treated with ALF523, ALF845 and ALF858 showed a mild course of the disease compared to vehicle treated rats even if these effects didn’t reach a statistical significance. In contrast, no effects were observed with the prophylactic treatment of all the tested compounds in the arthritis model induced in DBA mice by type II collagen (CIA). Some ALF compounds tested in both arthritis models (ALF821, ALF845, ALF523) that showed to ameliorate the clinical course of the AIA were not effective in the CIA model. This finding is not completely unexpected because different pharmacological responses between murine type II CIA and rat AIA have been reported and they may witness different immunopathogenetic mechanisms.
New CIA tests were being prepared at the time the project was prematurely suspended. Nevertheless, the rather encouraging results of the AIA tests and the variety of Mo(CO)4PyCa type complexes with different structural modifications that was already available encouraged the beginning of the study with the more advanced K/BxN model.
Altogether, 19 CORMs were tested in two AIA and one CIA study.

e) Start the evaluation of the anti inflammatory effect of the administration of a CORM in a mice model of arthritis (WP5).

A set of the previously selected CORMs which gave good indications in the AIA studies were judged appropriate to be tested in a more advanced animal model of arthritis, the K/BxN mice model. This is an interesting model to use in this work since it is a chronic model of arthritis. The tests were performed with five molecules arising from the previous tasks and their toxicity and anti-inflammatory properties were evaluated in vivo at the Instituto de Medicina Molecular (IMM). The molecules tested were ALF821, ALF826, ALF828, ALF843 and ALF844. Dexamethasone was used as a positive control in all experiments. The experiments were performed through the administration of each molecule in a daily dose of 30mg/Kg intraperitoneally, in groups of 5 mice. The administration of each molecule was maintained for a period of 36 days, and then the treatment was stopped for an additional period of 24 days after which all animals were sacrificed. This pause in the administration of CORMs allowed the study of arthritis relapse. The tests performed included blood analysis, histological analysis of the paws, as well as spleen and lymph nodes analysis in different time points throughout the 60 days of this study. The results showed that ALF826 was the molecule which more pronouncedly reduced the inflammatory score and delayed the appearance of symptoms after stopping the administration of the compound. ALF826 is also the CORM with higher drug efficiency. The results also revealed that ALF826 may exert its effects through the modulation of B cells and thus may possess a different immunopharmacological mode of action comparing with Dexamethasone. Moreover, it was also observed that ALF826 is able to decrease the production of the monocyte-specific chemokine 3 (MCP-3). Interestingly, this CORM was also the only molecule that was able to preserve the articular structure of front and hind paws, without any signs of inflammation, cellular infiltration and bone degradation, comparing with the other CORMs and also with Dexamethasone, as revealed by histological analysis.
These observations may suggest that ALF826 is a promising CORM molecule and maybe the best candidate for further studies.

f) Evaluation of the anti-inflammatory properties of the selected CORMs ex vivo in leukocytes isolated from RA patients (WP6)
After the evaluation of the anti-inflammatory effects of CORM administration in a mice model of arthritis described above, the best candidates selected were then tested ex vivo in primary cells isolated from blood samples collected from human RA patients in order to evaluate their anti-inflammatory properties at the Instituto de Medicina Molecular (IMM). To accomplish this main objective, the selected CORMs were tested ex vivo in RA patients’ leukocytes. The three best candidates selected in the previous tasks were assessed for their potential inhibitory effect on serum cytokine concentrations as well as on ROS production.
In RA it has already been reported that interleukin (IL)-1β, IL-6, IL-8, IL-17 and TNF have a high concentration in serum when compared to healthy donors, which correlates with increased disease activity. Thus, the first aim of this task was to evaluate the anti-inflammatory properties of the three new compounds in leukocytes isolated from blood of established RA patients, by quantifying the major pro-inflammatory cytokines (IL-1β, IL-6, IL-17, IL-22 and TNF) and chemokines (IL-8 and monocyte chemotactic protein 1, MCP-1) known to be involved in the pathogenesis of RA. For this, blood samples were obtained from 10 consecutive patients with established RA (more than 1 year of disease evolution), followed-up in the Rheumatology and Bone Metabolic Disease Department from Hospital de Santa Maria (HSM), to which the IMM is associated. These patients fulfilled all inclusion criteria for this study, namely having RA diagnosed according to the American College of Rheumatology (ACR) 1987 and were under DMARD treatment. Moreover, 10 healthy individuals were also recruited and used as controls. The study was approved by the ethical committee of HSM and held in accordance with the Declaration of Helsinki (as amended in Seoul 2008). Both controls and patients agreed to participate in this study by signing an informed consent form. Blood isolated leukocytes were incubated at 37°C in 5% of CO2 with the three compounds, ALF554, ALF854 and ALF826, at a concentration of 100 µM. Leukocytes were cultured with or without stimuli: phorbol myristate 12-13-acetate (PMA) and ionomycin. After 24h of culture, supernatants were collected for cytokine quantification and the viability of cells was quantified using Microtitration (MTT) test. The quantification of cytokines was performed by multiplex using the FlowCytomix technology according to the manufacturer’s instructions. Samples were acquired with a FACS Calibur flow cytometer and data was analysed by FlowCytomix Pro 2.4 software. The results obtained revealed that ALF826 can suppress IL-8 production in healthy control cells even after stimulation, and also that that ALF826 can diminish MCP-1 production in RA patients. MTT cell viability test carried on the leukocytes of healthy controls showed a survival rate always higher than 95%. On the other hand, leukocytes of patients with RA showed a survival rate below 95% for the following experimental conditions: ALF845 + stimulus (82 %); hepes + stimulus (81%); ALF554 (82%); ALF826 + stimulus (83%); DMSO (80%); DMSO + stimulus (75%). There was no statistical significant difference between the two study groups and between the different experimental conditions for the cytokines IL-1β, IL-6, IL-17, IL-22 and TNF. The results obtained for the chemokines IL-8 and MCP-1 revealed the potential anti-inflammatory properties of ALF826, which may act by modulating the state of activation of neutrophils and macrophages, the ones that arrive first at sites of inflammation and the two cell types that are present in a higher number in the synovial fluid and membrane of RA patients, respectively. Both monocytes/macrophages and neutrophils present in the serum and synovial fluid of RA patients contribute to an increase in the concentration of MCP-1 and IL-8, which influence their migration and are thought to play a central role in the inflammation and joint destruction characteristic of RA. Neutrophils also contribute to an increase of ROS.
The excessive production of ROS resulting from oxidative metabolism of leukocytes is associated with cell damage and increases in the inflammatory response. Therefore quantification of ROS production by leukocytes also provides important information. For this, leukocytes isolated from blood samples of the same experimental groups were cultured with 100 µM of each the three selected CORMs for 2h, and ROS production was measured using Burstest kit. Results of ROS species production by isolated leukocytes showed that ALF826 is capable of inhibiting the production of ROS in RA patients, thus demonstrating its putative anti-inflammatory activity, probably via the regulation of neutrophil activation (ROS production). Leukocytes of patients with RA showed a survival rate below 95% in the following experimental conditions: ALF554 + stimulus (72 %); DMSO + stimulus (78%).
Altogether, ALF826 seems to affect the activation and function of neutrophils and macrophages, and among the three new compounds analysed this maybe the best candidate for further studies.

g) Validation and Standardisation of Analytical Methods (WP7)
Participant Micron RS, together with the other participants in the project, selected two inflammation models set up by the CNR laboratory (Dr. Boraschi) to be standardised. The standardisation has begun with an in-depth analysis of the procedures currently used in Dr. Boraschi's laboratory and it was focused on the initial steps of both methods, namely Buffy Coat preparation, PBMC preparation, monocyte isolation and evaluation of monocyte concentration. Key parameters such as temperature, centrifugal conditions, cell characterisation criteria, cell viability acceptance criteria, reagents etc. were defined and established. For each step, a Standard Operating Procedure was issued and they represent the deliverables D7.1 and D7.2 for WP7 as indicated in the project.
In addition, Micron RS has prepared other Standard Operating Procedures regarding the quantification of human IL-1 and IL-6 from culture supernatants, the isolation of total RNA from cultured cells and the simultaneous quantification of a panel of human cytokines and chemokines in culture supernatants.
Micron RS, in collaboration with Participant 1 (Alfama), prepared two different protocols to standardise an assay, which was set up by Alfama laboratories, to determine the activity and cytotoxicity of CO releasing molecules (CORM). Among the different methods and assays, which were tested by Alfama scientists, the anti-apoptotic effect of CORMs on primary cultures of murine hepatocytes resulted in the best parameter for monitoring the activity of CORMs. Although murine hepatocytes are largely used for research purposes, their use to check the quality of any pharmaceutical preparation is very difficult. In fact cells isolated from normal liver (primary cultures of hepatocytes) show different histologic characteristics, therefore they are generally considered as a heterogeneous population of cells. In addition, the preparation of murine hepatocytes is not standardized, is long lasting and involves the use of animals. On the contrary, the use of established cell lines could be more feasible. In fact they are in general more homogenous, very well characterised and easy to culture. Alfama has also shown that certain CORMs have the ability to increase proliferation of liver cells (in vivo studies) or endothelial cells (in vitro). On the basis of these considerations, a panel of established cell lines, grown in pro-apoptotic conditions have been screened for their capability to proliferate in response to compounds ALF794 and ALF826. Two different apoptosis models were studied: serum starvation, and treatment with ActinomycinD (ActD) and Tumour Necrosis Factor (TNF-α). In initial experiments we tried to establish an apoptosis model by FBS starvation (0.5% FBS) in different cell lines. Unfortunately, the model did not work in the cell lines tested (PLC/PRF/5; HepG2; HCT116 and HT-29) and we were not able to induce apoptosis by FBS starvation. After the discussion of this protocol together with Alfama scientists, it was decided to try a different apoptosis model using ActinomycinD (ActD) and Tumour Necrosis Factor (TNF-α) to induce cell apoptosis.
The optimal concentration of ActD and TNF-α to induce approximately 50-60% decrease on cell survival has been studied for each cell line. This was investigated by a series of dose response experiments in the different cell lines prior to CORMs testing.
As soon as the optimal concentrations of ActD and TNF-α had been determined, the anti-apoptotic effect of ALF794, ALF826 and Hemin (positive control) was assessed in specific experiments.
The anti-apoptotic effect of the compounds on the cell lines was determined by using a proliferative/vitality assay (XTT).
The standardisation protocols of both models as well as the results of the experiments represent the deliverables of WP7.

h) Dissemination and Training (WP8)
The main dissemination activities put in place were
• Development of press releases, articles and features in local mass media,
• Presentation of the project´s results in various international biotech and life science meetings,
• website and electronic communication
Training activities were planned for the final workshop. Due to the premature ending of the project, this task was not achieved.

Potential Impact:
The current project was focused on a large unmet need, a treatment for rheumatoid arthritis, and on the development of a novel approach to meet that need, using CORMs. These compounds are potentially more effective, harmless and affordable than current solutions. Thus a novel but promising approach to RA management was explored, on rational basis, by SMEs leader in the field of CORMs. That fact in itself may represent a significant societal impact, albeit indirect and subject to the further development of the compounds, as RA represents an annual market of over €10 billion, causes about 40.000 deaths a year and has an inclaculable cost on the productivity of affected individuals and their communities. Any novel approach to the treatment of the disease has medical and social significance.
The participating SMEs were able to complement their core research capability, outsourcing research to RTD performers and research services providers. This activity strengthened the competitiveness of SMEs by enhancing their investment in RTD and by reinforcing their intellectual property position.
For the participating SMEs which hold the intellectual property and major knowledge on CORMs, especially Alfama, and on specific drug development technologies, this was a unique opportunity to sharply increase their value, by directly contributing for the placement of a of this kind near clinical testing. The collaborations at the European level added value to the work, improving the knowledge, science and technology that were shared among the various entities. Also, the visibility, diffusion and impact of the achievements will have been increased by the subsequent disseminations and exploitations plans of the projects, which took place at a European and global level as well.
At the beginning, the prospects of success were significant, due to the investment of critical mass and money in focused and planned Research and Technological Development activities. In the end the results came and were indeed very promising, with a clear reinforcement of the notion that CORMs had anti-arthritic effects at safe doses, and with the selection of at least one pre-clinical candidate – a molecule whose profile indicates it is a good candidate for clinical testing in this disease. Towards the end of the project the leading SME Alfama incurred in serious financial difficulties, as a consequence of challenges in fiding venture capital funding for its projects as a whole (which in turn were a consequence of the global financial crisis and of its particularly strong effect on countries such as Portugal). However, even if some activities could not be completed, the project’s central goals were achieved, and a major contribution was given to put forward a potential novel class of molecules – with a novel mechanism of action – for the treatment of RA.
The project also ended up contributing for the establishment of CO and CORMs as a novel therapeutic class, a fact consistent with the increasing number of academic groups using these compounds and number of peer-reviewed papers being published since the beginning of the project. Thus the beginning of a new sub-discipline of medicinal chemistry – based on the use of CO-releasing organometallic ompounds – may have happened, largely due to this project. This is important as an example of direct societal impact of the project, namely on the scientific and medical research community.

List of Websites:
www.altaweb.eu/cormra

Final Report Summary - CORMRA (Development of carbon monoxide releasing molecules for the treatment of rheumatoid arthritis)

Share this page

Download