*2.6. PGE2 Secretion*

Cell supernatants were collected from the 96-well plates used for the metabolic activity assay (MTT) after 24 and 72 h, and PGE2 secretion was analyzed. A commercial ELISA kit (Enzo Life Sciences, Farmingdale, NY, USA) was used to measure the amount (pg/mL) of PGE2 in the culture media, according to the manufacturer's instructions. The PGE2 concentration in each sample was determined following a previously reported procedure [30].

#### **3. Results and Discussion**

#### *3.1. Chemistry*

Compounds **1**-**9** were easily synthesized in good yields by reacting the terminal alkyne of propargylic intermediates and dicobalt(0)carbonyl octacarbonyl in tetrahydrofuran (THF). Detailed synthetic procedures are reported in the Supporting Information.

## *3.2. CO Release Assay*

The CO releasing behaviors of compounds **1**–**9** were evaluated through a myoglobin (Mb)-based spectrophotometric assay, considered the gold standard to analyze the CO releasing kinetics and a key criterion to select CORM structures [17,31]. This method analyzes the release of CO from CORMs by following the conversion of deoxymyoglobin (deoxy-Mb(II)) into CO-myoglobin (CO-Mb(II)) over time by UV-Vis spectroscopy. A 3.3 μM solution of compounds **1**–**9** was incubated with a 20 μM solution of deoxy-Mb (CORM/Mb 1:6 ratio), and a reducing agent (sodium dithionite) was added to prevent oxidation of deoxy-Mb(II) to Met-Mb(III). Changes in the absorption band in the Soret region of deoxy-Mb and Mb-CO were recorded every 30 min for 210 min, and a second derivative approach was applied to clearly discriminate between the three forms of Mb (deoxy-Mb, Mb-CO and Met-Mb). The relative amount of CO produced over time was calculated following a previously reported equation [23,32], and a correction factor was applied to account for Mb degradation induced by sodium dithionite. Concentration values of Mb-CO formed over time by compounds **1**–**9** are reported in Table 1. Their CO-release profiles are shown in Figure 2A–C, along with their *T*1/10 values, defined as the time necessary for a CO-RM to produce a concentration of CO-Mb of 1/10 of the initial (Figure 2D). The number of CO units released by CORMs **1**-**9** after 210 min is reported in Table 2. For selected compounds, the assay was performed using a 1:1 CORM/Mb ratio (Figure 3A–D). The aim was to explore their ability to release CO in a less favored condition, considering that CO release from CORMs is stimulated by an excess of Mb.


**Table 1.** MbCO formed at each time point when compounds **1**–**9** were analyzed at 1:6 CORM:Mb ratio.

**Table 2.** CO units released by compounds **1**–**9** after 210 min of incubation working at 1:6 CORM-Mb ratio and by compounds **1**, **5** and **6** at 1:1 CORM-Mb ratio.


**Figure 2.** (**A**–**C**): CO-release profiles of compounds **1**–**9** as Mb-CO formed over time (1:6 CORM:Mb ratio); (**D**): *T*1/10 values for compounds **1**–**4** and **6** (defined as the time necessary for a CORM solution to produce a Mb-CO concentration of 1/10 of the initial).

All the tested compounds were effective CO releasers, with CO release kinetics comparable to previously reported DCH-CORMs (slow CO release up to 210 min) [22,23,33]. The obtained data showed considerable differences among the analyzed compounds, depending on the heterocyclic nucleus and the different alkyne system linked to the DCH group (Table 1).

As shown in Figure 1, compounds **1**–**5** bear a propynyl-pyrrol-3-yl-acetate moiety, compound **6** a propynyl-pyrrole-2-carboxylate motif and compound **7** a propynyl benzoate group at position 1 of the pyrrole ring. Compounds **8** and **9** are characterized by a 3-((prop-2-yn-1-yloxy)methyl)-1*H*-pyrazole scaffold. Derivatives **1**–**5** showed comparable kinetics, with a sustained release of CO over time (Figure 2A). The kinetics of release were almost superimposable over the first 120 min, with negligible differences over the last 90 min. Despite slight differences in their *T*1/10 (Figure 2D), the comparison of the units of CO released over time (Table 2) displayed a very similar behavior within the series, suggesting that the substitution pattern on the aryl rings only slightly impacts their CO-releasing abilities. In particular, compounds **1**, **2** and **5**, bearing a sulfonyl group on the aryl ring at C5, released almost the same amount of CO at the end of the assay (0.64–0.68 CO units, Table 2). Compounds **3** and **4**, both decorated with a sulfamoylphenyl moiety at position 1 of the pyrrole, displayed very similar behavior and released a comparable amount of CO after 210 min (0.74–0.75 CO units, Table 2). Pyrrole **6** showed a completely different CO releasing profile (Figure 2B,C) and was the fastest and most efficient CO-releaser of the series, with a *T*1/10 value of 99.01 min (Figure 2D) and 0.92 CO units released after 210 min (Table 2). This compound showed a fast release of CO over the first 120 min, which reaches a maximum at 150 min and then slows down. When compared to its analogue **1**, compound **6** produced a 1.35-fold higher amount of Mb-CO at each time point until the end of the assay. These data suggest that the group bearing the DCH moiety and the chemical space around it strongly influences the CO release kinetic. The

CO releasing behavior of compound **7** supported this hypothesis: indeed, it releases CO slower than compounds **1**-**5**, although reaching almost the same amount of Mb-CO after 210 min of incubation (Figure 2B,C). Differently from compounds **1**-**6**, the DCH portion of this compound is linked to a propynyl benzoate group at position 1 of the pyrrole ring, suggesting that the electronic structure can induce different CO releasing properties. Pyrazoles **8** and **9** showed different CO releasing profiles (Figure 2C). At the end of the assay, these compounds produced much smaller values of Mb-CO (1.75 μM and 1.22 μM for **8** and **9**, respectively, Table 1) than compounds **1**–**7** at the same time point. Therefore, it is interesting to note that the 3-((prop-2-yn-1-yloxy)-methyl)-1-pyrazole moiety is probably detrimental in terms of CO releasing efficiency when compared to acetate, carboxylate or benzoate moieties decorating compounds **1**–**7**. Previous studies reported the impact of the group bearing the DCH moiety in determining the CO releasing properties [22,23]. The different CO release kinetics observed for compounds **1**–**9** confirm the influence of the drug sphere on CO releasing properties and suggest that both the electronic density around the pyrrole/pyrazole ring and the group bearing the DCH moiety strongly impact the rate of CO release, yet further studies are needed to better characterize this phenomenon. To further explore the releasing properties of these derivatives, we selected compounds **1**, **5** and **6** to be studied at different CO-RM:Mb ratios (Figure 3).

**Figure 3.** CO release profiles of compounds **1** (**A**), **5** (**B**) and **6** (**C**) analyzed at 1:6 (filled square) and 1:1 (empty triangles) CORM-Mb ratios; (**D**): CO units released by compounds **1**, **5** and **6** after 210 min of incubation working at 1:6 (blue columns) and 1:1 (orange columns) CORM-Mb ratios.

As mentioned above, an excess of the CO acceptor (Mb) stimulates the CO release from CORMs. Thus, we expected a decrease in CO release at 1:1 CO-RM:Mb ratio. As shown in Figure 3, after 210 min of incubation, compounds **1**, **5** and **6** released a lower amount of CO when compared to the one observed in 1:6 conditions (Figure 2). Moreover, all the compounds released the same CO units (0.11–0.12 CO units), regardless of their different chemical structure. Therefore, disfavoring the CO release seems to reduce the differences in CO releasing efficiencies observed when Mb is present in excess (1:6/CO-RM:Mb).

#### *3.3. Effects of Compounds* **1**–**5** *on Human Tenocytes*

Once the influence of electronic and steric properties on CO release has been established, a fine-tuning of the drug sphere should also focus on treating particular conditions and on the biological activity of the drug sphere itself [17,34]. Compounds **1**-**5** drug sphere belongs to a series of sulphone and sulfamoyl diarylpyrrole derivatives developed by our research group as COX-2 selective inhibitors [35–39]. This class of compounds showed promising in vitro and in vivo anti-nociceptive and anti-inflammatory properties and tolerates a wide range of substituents at position C3. As augmented PGE2 levels are a marker of oxidative-stress inflammation, the modulation of PGE2 secretion might be a valuable strategy for therapeutic intervention of tendon diseases [27,40]. We therefore speculated that the conjugation of a COX-2 inhibiting scaffold and a CO-releasing moiety could help to achieve promising CORM-candidates for the treatment of tendon inflammatory-based diseases, as COX-2 inhibition and CO release could act synergically to resolve inflammation and restore oxidative homeostasis. Moreover, the conjugation of structural fragments of anti-inflammatory drugs with metal carbonyl moieties is well documented in the literature [18,19,24], and the five selected compounds showed quite similar CO release profiles (Figure 2), allowing us to make a proper comparison and rationalization of the observed biological activities. To help to discriminate between COX-2 mediated and independent activities, the efficacies of these compounds against inflammation and oxidative cytotoxicity were studied through the analysis of different parameters: the metabolic activity of tenocytes before and after H2O2 stimulation and the quantification of PGE2 secretion.

Tendinopathies are characterized by a higher level of tenocyte apoptosis and a decreased metabolic activity, which can reduce the resistance of tendon structures and lead to failure in healing [41,42]. Unstimulated tenocytes were therefore exposed to increasing concentrations of CORMs to evaluate their biocompatibility and effects under non-oxidative stress conditions (Figure 4). It is worth noting that Meloxicam exerted no significant effects on tenocytes when administered in the same experimental conditions [26].

On the other hand, compounds **1**–**5** significantly increased the metabolic activity of tendon-derived cells after 24 h, as observed for CAI-CORMs hybrids [26]. In more detail, all the tested compounds showed a dose-dependent rise up to 25–50 μM already after 24 h of exposure, which was particularly significant in the presence of compounds **3** and **5**. Interestingly, the tested compounds seemed less active after a 72 h exposure, as the metabolic activity was comparable to that of the control up to the concentration of 25 μM. This might be related to their slightly different CO release kinetics, but further studies are needed to corroborate this hypothesis.

#### *3.4. Establishment of the Inflammatory Cell Model and Effects of Compounds* **1**–**5** *on Human Tenocytes under Oxidative Stress Conditions*

These preliminary results highlighted that compounds **1**–**5** have good proliferative effects on tendon-derived cells and provided a proof-of-concept that the biological features of these compounds are not only COX-2 mediated but also rely on CO release. As reported elsewhere [43], increasing cell metabolism and proliferation are particularly important for tendon tissue repair after the acute inflammatory phase. With this rationale, compounds **1**–**5** were tested in sub-toxic oxidative stress conditions in vitro [26] to investigate their ability to counteract H2O2-induced oxidative stress. After 3 h of incubation with 100 μM H2O2, human tenocytes were exposed to increasing concentrations of CORMs. As reported in Figure 5, all the tested compounds were more active than Meloxicam in increasing the cell metabolism of tenocytes. Notably, compounds **3**–**5** were the most active of the series. In particular, compound **3** showed an outstanding efficacy, being able to increase metabolic activity up to 166.2% after 72 h when administered at 25 μM. Unlike under non-oxidative stress conditions, the percentage of metabolically active tenocytes increases after 72 h. This observation suggests a composite mechanism of action, which

probably results from the combination of COX-2 inhibition and CO release. Consistent with previously obtained results, activities were maximum at 25 μM, then decreased at the higher concentrations tested.

**Figure 4.** Metabolic activity of human primary tendon-derived cells exposed to increasing concentrations of CORMs (compounds **1**–**5**) after 24 and 72 h. The control sample (0 μM = cells treated with DMSO 0.1%) is set as 100%. a = *p* < 0.01; b = *p* < 0.001; c = *p* < 0.0001 between cells treated with CORMs and the control sample.

**Figure 5.** Metabolic activity of H2O2-pre-incubated human primary tendon-derived cells exposed to increasing concentrations of CORMs (compounds **1**–**5**) after 24 and 72 h. Cells were pre-incubated with H2O2 100 μM for 3 h. The control sample (0 μM = cells pre-incubated with H2O2 and treated with DMSO 0.1%) is set as 100%. a = *p* < 0.01; b = *p* < 0.001; c = *p* < 0.0001 between cells treated with CORMs and the control sample.
