PEG Linker Length Strongly Affects Tumor Cell Killing by PEGylated Carbonic Anhydrase Inhibitors in Hypoxic Carcinomas Expressing Carbonic Anhydrase IX

Hypoxic tumors overexpress membrane-bound isozymes of carbonic anhydrase (CA) CA IX and CA XII, which play key roles in tumor pH homeostasis under hypoxia. Selective inhibition of these CA isozymes has the potential to generate pH imbalances that can lead to tumor cell death. Since these isozymes are dimeric, we designed a series of bifunctional PEGylated CA inhibitors (CAIs) through the attachment of our preoptimized CAI warhead 1,3,4-thiadiazole-2-sulfonamide to polyethylene glycol (PEG) backbones with lengths ranging from 1 KDa to 20 KDa via a succinyl linker. A detailed structure−thermal properties and structure–biological activity relationship study was conducted via differential scanning calorimetry (DSC) and via viability testing in 2D and 3D (tumor spheroids) cancer cell models, either CA IX positive (HT-29 colon cancer, MDA-MB 231 breast cancer, and SKOV-3 ovarian cancer) or CA IX negative (NCI-H23 lung cancer). We identified PEGylated CAIs DTP1K 28, DTP2K 23, and DTP3.4K 29, bearing short and medium PEG backbones, as the most efficient conjugates under both normoxic and hypoxic conditions, and in the tumor spheroid models. PEGylated CAIs did not affect the cell viability of CA IX-negative NCI-H23 tumor spheroids, thus confirming a CA IX-mediated cell killing for these potential anticancer agents.


Introduction
Genetic and epigenetic mutations can transform normal human cells into malignant ones. As the human body ages and normal repair mechanisms recede, the risk of cancer increases, making it currently the second leading cause of death worldwide. The malignant cells lose their normal physiologic functions and contact inhibition and gain characteristics that facilitate their survival, growth, and metastasis [1,2]. Cancer cells grow aggressively, forming tumors in different parts of the body. Often, their aggressive growth outpaces the growth of supporting vasculature, so they become hypoxic [1][2][3][4]. Many aggressive tumors such as breast, colorectal, ovarian, pancreatic, bladder, and other carcinomas are associated with hypoxia and the upregulation and stabilization of hypoxia-inducible factor 1 (HIF-1) [1][2][3][4][5][6][7]. Subsequently, HIF-1 relocates to the nucleus, where it triggers the expression of many proteins required for malignant tumor survival under hypoxic conditions [5][6][7]. Among them are the isozymes of carbonic anhydrase (CA), namely CA IX and CA XII [8][9][10][11].
Carbonic anhydrases (CAs) are metalloenzymes ubiquitously present in both prokaryotes and eukaryotes, where they serve as catalysts for the reversible hydration of CO 2 to Carbonic anhydrases (CAs) are metalloenzymes ubiquitously present in both prokaryotes and eukaryotes, where they serve as catalysts for the reversible hydration of CO2 to bicarbonate and a proton. Eight separate gene families encode for these enzymes: α-, β-, γ-, δ-, ζ-, η-, θ-, and ι-CAs [8,12]. The α-CAs are the only group found in mammals, with 14 different CA isozymes being described in humans (hCAs) [8,12]. Of these hCAs, two isozymes are mitochondrial (CA VA and CA VB), five isozymes are present in the cytosol (CA I, CA II, CA III, CA VII, and CA XIII), four isozymes are membrane-bound (CA IV, CA IX, CA XII, and CA XIV), and one isozyme is secreted in saliva and milk (CA VI) [8,12]. Through these isozymes, CA plays an important role in respiration and CO2 excretion, pH homeostasis, secretion of electrolytes, gluconeogenesis, lipogenesis and ureagenesis, bone resorption and calcification, and tumorigenicity [8,12]. Focusing on tumorigenicity, it must be emphasized that CA IX and CA XII play central roles in hypoxic tumor cell biochemistry ( Figure 1) [6,8,9,13]. A lack of oxygen triggers upregulation of glycolysis as the main source of ATP production, with pyruvate being converted to lactate to regenerate NAD + . Glycolysis produces protons which are combined with intracellular HCO3 − to produce CO2 and H2O in a reaction catalyzed by cytosolic CAs, especially CA I and CA II. Carbon dioxide passes through the cell membrane, and it is rehydrated outside the cell by overexpressed CA IX and CA XII isozymes to generate HCO3 − and H + . The bicarbonate ion is imported back into the cell and exchanged with Cl − via anion exchangers. The tandem action of cytosolic isozymes with the overexpressed membrane-bound CA IX and CA XII allows the efficient transport of cytosolic protons formed in massive amounts from upregulated glycolysis from inside the tumor cell to the extracellular milieu in an ATPindependent manner. Thus, the pH of the cytoplasm of tumor cells is maintained within normal limits while acidifying the extracellular environment, with detrimental consequences for the normal cells surrounding the tumor (Figure 1) [6,9,13].  glycolysis is significantly upregulated in tumor cells to make ATP under hypoxic conditions. Protons produced via glycolysis are transported outside the cell via various mechanisms, with cytosolic (CA I and CA II) and membrane-bound CA isozymes (CA IX and CA XII over-expressed in hypoxic tumors) playing a central role. Interfering with tumor pH homeostasis through the inhibition of CA IX and CA XII thus has the potential to control and suppress the proliferation of cancer cells.
Building on these premises and taking into account that CA IX and XII are dimers in vivo [50,51], our group developed a series of bifunctional PEGylated CAIs 20-25 by attaching known CAI pharmacophores (aminobenzenesulfonamide, 5-amino-1,3,4-thiadiazole-2-sulfonamide, and aminobenzolamide) on oligoethylene glycol (EG) or on a PEG2000 polymeric backbone (PEG2K) (Scheme 1) [52] All the bifunctional PEGylated and oligoethylene glycol analog CAIs 20-25 were tested for their ability to inhibit off-target cytosolic and membrane-bound isozymes of carbonic anhydrase, including CA IX and CA XII. The results ( Table 1) showed an excellent inhibition profile against membrane-bound isozymes CA IX and CA XII, with nanomolar potency of most representatives against these targets. [52]. Interestingly, the long PEG linker increased the selectivity of bis-sulfonamides 21, 23, and 25 for the membrane-bound isozymes versus the cytosolic ones (especially against CA II) compared with their oligoethylene glycol congeners 20, 22, and 24 (Table 1). Subsequent in vitro testing of oligo/polyEG CAIs 20-25 was performed in 2D under normoxic and hypoxic conditions and in 3D (tumor spheroids) models of cancer. The CAIs were assessed for their ability to reduce the viability of the colon carcinoma HT-29 cell line, the breast cancer MDA-MB 231 cell line, and the ovarian cancer SKOV3 cell line, all expressing CA IX and CA XII isozymes [52][53][54][55][56][57][58][59][60][61][62][63][64][65][66][67][68][69]. The in vitro testing identified PEGylated bis-sulfonamide 23, bearing the classical 1,3,4-thiadiazole-2-sulfonamide "warhead", as the most efficient inhibitor, which displayed consistent and significant cancer cell killing at concentrations of 10-100 µM across all three cell lines in 2D/3D cellular models under both normoxic and hypoxic conditions [52]. We assigned this excellent potency against purified CA IX and CA XII isozymes and the consistent and significant cancer cell killing obtained with PEGylated CAI 23 to the cooperative binding and bivalent association of this type of two-pronged inhibitor with the two neighboring active sites [70], as shown by Whitesides [71].
Based on this cooperative binding mechanism, it is expected that PEGylated bissulfonamide congeners of 23 having different PEG backbone lengths will display the 1,3,4-thiadiazole-2-sulfonamide warhead differently and will inhibit the dimeric CA IX and CA XII in a different manner. In order to test our working hypothesis, we designed five different PEGylated bis-sulfonamides bearing PEG 1K, PEG 2K, PEG 3.4K, PEG 5K, and PEG 20K polymeric backbones, which were synthesized via our standard procedure and were tested in vitro for their ability to kill cancer cells expressing CA IX and CA XII.

Synthesis of Polymeric CAIs
The new series of compounds 23 and 28-31 were synthesized following previously reported procedure (Scheme 2) [52]. The starting 5-amino-1,3,4-thiadiazole-2-sulfon-amide 26 was synthesized as previously described [31,72,73]. The succinyl derivatives 27 was generated from the condensation of 5-amino-1,3,4-thiadiazole-2-sulfonamide 26 with succinic anhydride in acetonitrile at room temperature. The carboxy group of succinylamido sulfonamide was subsequently activated with 2-chloro-4,6-dimethoxy-1,3,5-triazine (CDMT)/Nmethylmorpholine (NMM) [52,74,75] and was condensed with diamino polyethylene glycols having different PEG lengths to yield the bis-sulfonamides 28-31. These compounds were purified by flash chromatography and characterized by standard analytical methods (see the Materials and Methods section). Based on this cooperative binding mechanism, it is expected tha fonamide congeners of 23 having different PEG backbone lengths w thiadiazole-2-sulfonamide warhead differently and will inhibit the di XII in a different manner. In order to test our working hypothesis, w ferent PEGylated bis-sulfonamides bearing PEG 1K, PEG 2K, PEG 3.4 20K polymeric backbones, which were synthesized via our standard tested in vitro for their ability to kill cancer cells expressing CA IX an

Scheme 2.
Synthetic strategy for the generation of bis-sulfonamide CA inhib

DSC Analysis of Polymeric CAIs
Differential scanning calorimetry (DSC) is a powerful method to polymers and polymer conjugates and to characterize the thermal beh Scheme 2. Synthetic strategy for the generation of bis-sulfonamide CA inhibitors.

DSC Analysis of Polymeric CAIs
Differential scanning calorimetry (DSC) is a powerful method to assess the purity of polymers and polymer conjugates and to characterize the thermal behavior of these chemical entities in bulk, which are important for pharmaceutical formulations containing them [76]. Consequently, DSC was used to analyze the thermal properties of the new PEGylated bis-sulfonamides DTP2K 23 and DTP1K 28-DTP20K 31 ( Figure 2). DSC thermograms of polyethylene glycol 1000 (PEG-1K-OH) and polyethylene glycol 1000 diamine (PEG-1K-NH2) were also presented for comparison.  DSC thermograms from Figure 2a in the first cooling cycle and second heating cycl (5 °C/min heating/cooling rate, respectively) revealed that polyethylene glycol 1000 (PEG 1K-OH) has a sharp phase transition with a crystallization temperature Tc of 29.30 °C and a melting temperature Tm of 39.18 °C. The approximately 10 °C supercooling reflects relatively easy adoption of a crystalline structure due to the high mobility -OH small end group. A change of the end functional group from -OH to -NH2 in diamino PEG (PEG 1K-NH2) slightly affected the stability of the PEG backbone, which presented a Tc at lowe temperature (23.40 °C). The Tm (37 °C) was also lower than the one observed for PEG-1K OH, revealing that amino groups slightly destabilize the PEG backbone. The lower stabil ity of PEG-1K-NH2 compared with PEG-1K-OH was also reflected by the lower enthalp observed during crystallization/melting (≈168 J/g for PEG-1K-NH2 vs. ≈202 J/g for PEG 1K-OH). The conjugation of warhead 2-sulfonamide-1,3,4-thiadiazole-5-(succinamido) on PEG significantly destabilized the PEG backbone. PEGylated bis-sulfonamide DTP1K DSC thermograms from Figure 2a in the first cooling cycle and second heating cycle (5 • C/min heating/cooling rate, respectively) revealed that polyethylene glycol 1000 (PEG-1K-OH) has a sharp phase transition with a crystallization temperature Tc of 29.30 • C and a melting temperature Tm of 39.18 • C. The approximately 10 • C supercooling reflects a relatively easy adoption of a crystalline structure due to the high mobility -OH small end group. A change of the end functional group from -OH to -NH2 in diamino PEG (PEG-1K-NH2) slightly affected the stability of the PEG backbone, which presented a Tc at lower temperature (23.40 • C). The Tm (37 • C) was also lower than the one observed for PEG-1K-OH, revealing that amino groups slightly destabilize the PEG backbone. The lower stability of PEG-1K-NH2 compared with PEG-1K-OH was also reflected by the lower enthalpy observed during crystallization/melting (≈168 J/g for PEG-1K-NH2 vs. ≈202 J/g for PEG-1K-OH). The conjugation of warhead 2-sulfonamide-1,3,4-thiadiazole-5-(succinamido)-on PEG significantly destabilized the PEG backbone. PEGylated bissulfonamide DTP1K 28 did not show any phase transition in the first cooling cycle. Upon heating, the glassy material crystallizes at approximately −18 • C and then melted at around 23 • C. The enthalpy observed upon melting was about 1/3 of the corresponding thermal effect observed for PEG-1K-NH2, reflecting the strong destabilizing effect of the bulky warhead on the (relatively short) PEG backbone. Elongation of the PEG backbone (Figure 2b) allowed better organization of the PEG linker upon cooling. DTP2K 23 managed to crystallize upon cooling (with about 30 • C supercooling). Longer PEG linkers (3.4K to 20K) raised the Tc and reduced the supercooling effect monotonously, from ≈17 • C (DTP3.4K 29) to ≈16.3 • C (DTP5K 30) to ≈14.4 • C (DTP20K 31). It appears that once the linker length reached a critical length (≈3.4K), the impact of the terminal warheads on the backbone is minimized and the two terminal warheads can be considered isolated from one another. Tm increased monotonously from 36.93 • C (DTP3.4K 29) to 58.81 • C (DTP20K 31), as expected. Interestingly, the crystallization/melting enthalpy did not follow a linear trend. DTP3.4K 29 displayed the highest thermal effect (≈142 J/g), followed by DTP20K 31 (≈140 J/g) and DTP5K 30 having an intermediate value (≈118 J/g). This trend probably reflects the different relative positions of the two warheads and the impact of PEG length and coiling towards their separation. We postulate that the two warheads are relatively well separated in DTP3.4K and DTP20K and can come closer to each other in DTP5K, thus hampering crystallization and reducing the enthalpy associated with it.
In the 2D cell viability assay, cells were grown sub-confluent in Petri dishes and then plated in 96-well plates in normal conditions (37 • C, 5% CO 2 in air). For each cell line, half of the plates were incubated in normal (normoxic) conditions, while the other half were subjected to a hypoxic condition (1% O 2 , 5% CO 2 , and 94% N 2 ) to induce the expression of the CA IX isozyme. Subsequently, the plates were treated with CAI solutions in media for 24 h of incubation time at three different concentrations (1 mM, 100 µM, and 10 µM) in either normoxic/hypoxic conditions. Cell viability was measured using an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Figure 3b,c) [77].
The data from Figure 3 show that the ability of the new PEGylated CAIs to kill the tumor cells correlated well with CA IX expression. All PEGylated CAIs affected tumor cell viability in CA IX expressing cancer cell lines, with a greater effect observed under hypoxic conditions, where CA IX is overexpressed.
Under normoxic conditions, the viability of the HT-29 cell line was significantly reduced after treatment with bis-sulfonamides DTP2K 23 and DTP1K 28-DTP20K 31, with CAI having shorter PEG backbones being more efficient. Bis-sulfonamide DTP1K 28 reduced the cell viability to around 45% and was the best inhibitor in these conditions for this cell line. The order of cell killing efficiency was DTP1K 28 > DTP2K 23 > DTP3.4K 29 > DTP5K 30 > DTP20K 31 > AAZ 1. The breast cancer MDA-MB231 and ovarian cancer SKOV-3 cell lines were less sensitive to bis-sulfonamide treatment under normoxic conditions, with DTP1K 28 and DTP2K 23 being again the most efficient, with an effect matching that of AAZ 1. Likewise, the viability of CA IX-negative cell line NCI-H23 was not affected after treatment with these CAIs, implying a CA IX-mediated cell killing for these inhibitors.   However, under hypoxic condition, all three CA IX expressing cell lines were significantly affected by the CAI treatment in a concentration-dependent manner, as expected, with the colon HT-29 and breast MDA-MB231 cancer cells being slightly more sensitive to CAI treatment than ovarian cancer cell line SKOV-3. All the PEGylated CAIs DTP2K 23 and DTP1K 28-DTP20K 31 showed significant tumor cell killing abilities. The cell killing efficiency decreased as the length of the PEG linker increased. PEGylated CAIs DTP1K 28, DTP2K 23, and DTP3.4K 29, bearing short and medium PEG backbones, were the most efficient in all three CA IX expressing cell models. The PEGylated bis-sulfonamide DTP1K 28 was found to be the best inhibitor against all three cell lines under hypoxic conditions, followed by DTP2K 23 and DTP3.4K 29. Both DTP1K 28 and DTP2K 23 displayed robust cell killing, with cell viabilities as low as 30-40% at 1 mM, 50% at 100 µM, and 70-80% at 10 µM, being influenced by the cell type and presence of CA IX in large amounts due to hypoxia. On the other hand, inhibitors DTP5K 30 and DTP20K 31, having long PEG linkers, were generally less efficient, even at high concentrations. Inhibitor All the PEGylated CAIs DTP2K 23 and DTP1K 28-DTP20K 31 showed superior cell killing profiles to the positive control AAZ 1 used in this study, which was found to be less efficient toward tumor cell killings under both hypoxic and normoxic conditions. AAZ 1 was efficient only at very high concentrations, thus revealing the high impact of the PEG backbone on the efficiency of cell killing. Importantly, the CA IX-negative cell line NCI-H23 did not respond to the treatment with these PEGylated CAIs, even under hypoxic conditions.
The cell viability data shown in Figure 4 correlated well with the level of CA IX expression in tumor spheroids (Figure 3a), similarly to the 2D cell cultures (Figure 3b,c). Tumor spheroids become hypoxic due to their 3D growth that limits oxygen perfusion into the spheroid cores. Therefore, the tumor spheroid model is closer to the in vivo conditions. The higher level of expression of CA IX in HT-29 and SKOV-3 tumor spheroids allows one to establish the structure-activity relationship (SAR) for the new PEGylated CAIs even at low inhibitor concentrations of 100 µM and 10 µM. Among the three CA IX expressing cell line, HT-29 and SKOV-3 tumor spheroids were found to be most sensitive to bis-sulfonamide CAI treatments, in good correlation with the CA IX expression level, with the MDA-MB231 cell line being slightly less affected. All the PEGylated CAIs impacted the viability of tumor spheroids for all three CA IX expressing 3D cell cultures, with cell killing efficiency decreasing with an increase in PEG linker length. At high concentrations, all CAIs were efficient in reducing cell viability. At lower concentrations, the shorter CAIs were again more efficient. Thus, PEGylated bis-sulfonamide CAIs having shorter and medium PEG linker DTP1K 28, DTP2K 23, and DTP3.4K 29 were the most efficient CAIs in this assay. PEGylated CAI DTP1K 28 at 1 mM/100 µM/10 µM significantly decreased the cell viability to about 30%/68%/72% in HT-29, 40%/50%/86% in SKOV-3, and 55%/70%/83% in MDA-MB231, whereas for DTP2K 23 treatment at the respective concentrations, cell viability decreased to about 50%/75%/80% in HT-29, 55%/85%/100% in SKOV-3, and 45%/62%/64% in MDA-MB 231. High molecular weight PEGylated CAIs DTP5K 30 and DTP20K 31 were efficient in killing the cells only at higher concentrations. At 1 mM, CAI DTP5K 30 decreased the cell viability to 49% in HT-29, 57% in SKOV-3 and to 61% in MDA-MB231 tumor spheroids, while PEGylated CAI DTP20K 31 only significantly impacted the viability of HT-29 tumor spheroids, expressing the highest level of CA IX. Thus, the length of the PEG linker plays a significant role in killing tumor spheroid cells, with short PEG 1K and 2K linkers being the most efficient for this type of CAIs. Again, all the new PEGylated CAIs exhibited better cell killing efficiency than the positive control AAZ 1, which was found to reduce cell viability to about 80% (for 1 mM) and was ineffective at lower concentrations. The cell viability data shown in Figure 4 correlated well with the level pression in tumor spheroids (Figure 3a), similarly to the 2D cell cultures ( Tumor spheroids become hypoxic due to their 3D growth that limits oxyge Importantly, PEGylated CAIs DTP2K 23 and DTP1K 28-DTP20K 31 did not affect the cell viability of CA IX-negative NCI-H23 tumor spheroids, thus reconfirming CA IX-mediated cell killing by these inhibitors. These data indicate that the length of the PEG backbone plays an important role in tumor cell killing, with the PEG linker significantly affecting the availability of sulfonamide warhead groups for binding into the active sites of dimeric membrane bound CA IX isozyme. Considering a proposed cooperative binding of the bifunctional inhibitor to the two active sites and the fact that PEG may adopt a coiled structure, it is conceivable that a longer linker might place one warhead towards the interior of the coiled backbone, thus decreasing its availability in binding to the dimeric CA IX and therefore lowering the potency, both in vitro and in vivo. Based on the crystal structure of the dimeric CA IX [51], one can calculate that even the PEG 1K is long enough to allow positioning of the two warheads of the PEGylated bis-sulfonamide in the two individual active sites of the CA IX dimer. This cooperative mechanism of action may be different from the interaction of CA IX with classical monodentate inhibitors [8,26,79,80] and warrants further investigations. Techniques. The purity and structural identity of the intermediates and of the final products were assessed by a combination of techniques that included thin-layer chromatography (TLC), HPLC-MS, 1 H-NMR, COSY, 13 C-NMR, high resolution mass spectrometry (HR-MS), and gel permeation chromatography (GPC). TLC was carried out on SiO 2precoated aluminum plates (silica gel with F254 indicator; layer thickness 200 µm; pore size 60 Å, from Sigma-Aldrich (St. Louis, MO, USA)).

Materials
The purity and the structure identity of the intermediates and of the final products were assessed by a combination of techniques that included thin-layer chromatography (TLC), HPLC-MS, 1 H-NMR, COSY, 13 C-NMR, high resolution mass spectrometry (HR-MS), and gel permeation chromatography (GPC). TLC was carried out on SiO 2 -precoated aluminum plates (silica gel with F254 indicator; layer thickness 200 µm; pore size 60 Å, from Sigma-Aldrich).
NMR spectra were recorded at ≈300 K with a Bruker Avance III 400 Plus spectrometer equipped with a 5-mm indirect detection probe operating at 400 MHz for 1 H-NMR and at 100 MHz for 13 C-NMR. Chemical shifts are reported as δ values, using tetramethylsilane (TMS) as an internal standard for proton spectra and the solvent resonance for carbon spectra. Assignments were made based on chemical shifts, signal intensity, COSY, HMQC, and HMBC sequences. For 1 H-NMR, data are reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, t = triplet, sep = septet, and m = multiplet), coupling constants J (Hz), and integration.
Differential scanning calorimetry (DSC) was performed using a TA Instruments Q200 MDSC (New Castle, DE, UK) and a heating/cooling rate of 5 • C/min for all compounds.

General Procedure for the Preparation of Bis-Sulfonamides 23 and 28-31
In a flame-dried 20-mL reaction vial, succinylamidosulfonamide 27 (0.536 g, 1.92 mmol) and 2-chloro-4,6-dimethoxy-1,3,5-triazine (CDMT, 0.37 g, 2.11 mmol) were dissolved in 2 mL dry DMF. The reaction vial was cooled in an ice bath at 0 • C. After 5 min, Nmethylmorpholine (NMM, 235 µL, 0.213 g, 2.11 mmol) was added and the reaction mixture was stirred at 0 • C for 45 min under nitrogen atmosphere. Subsequently, diamino PEG (H 2 N-PEG-NH 2 , 0.500 g, 0.250 mmol), dissolved in 1.5 mL dry DMF, was added dropwise to the reaction mixture. The reaction was stirred at room temperature under nitrogen atmosphere for two days. The DMF was evaporated by rotavap under high vacuum, and the crude compound was absorbed on SiO 2 . Flash chromatography was performed with CHCl 3 /MeOH gradients on a SiO 2 column. Useful fractions (determined by TLC, performed with DCM/MeOH (75/25 v/v), and by NMR) were grouped and evaporated to dryness to yield a relatively pure compound which was further purified by another flash chromatography run. Finally, the pure compounds were dried under high vacuum for 24 h.

Viability Evaluation in 2D Cell Cultures
Four cell lines, namely HT-29, MDA-MB-231, SKOV3, and NCI-H23, were cultured at 37 • C and 5% CO 2 using RPMI-1640, Dulbecco's Modified Eagle's medium (DMEM), McCoy's 5A Medium, and RPMI-1640, respectively, supplemented with 10% fetal bovine serum. When the cells were about 75% confluent in 75-cm 2 flasks, they were trypsinized with trypsin-EDTA, counted, and plated in 96-well plates at a density of 10 4 cells/well with a final volume of 200 µL in each well. Two plates were made for each cell line. After 24 h of incubation in a 5% humidified incubator at 37 • C, one plate from each cell line was placed in the hypoxia chamber and purged for 10 min with the hypoxia gas mixture (1% O 2 , 5% CO 2 , balance N 2 ). After 1 h, the chamber was purged again with hypoxic gas mixture and placed back in the incubator. After a 24-h incubation in the incubator at 37 • C, both the normoxic and hypoxic plates were retrieved and the media were removed from all wells. Cells were washed with 5% sterile dextrose and treated with different solutions of different concentrations of CAIs DTP2K 23 and DTP1K 28-DTP20K 31 in media, all pre-filtered through a 0.2-µm sterile nylon filter. Each assay was done at least in triplicate. After 24 h of incubation time with CAIs, the supernatant was discarded; cells were washed with phosphate-buffered saline (PBS); and after the removal of PBS, each well received 120 µL of MTT/media solution (prepared from 5 mg/mL MTT in sterile PBS, added to media in the ratio 1:6). The plates were incubated for 4 h at 37 • C, the MTT/media was aspirated off, and 150 µL of DMSO was added to each well to dissolve the formazan crystals. The plates were stored in the incubator for 5 min at 37 • C, and absorbance was read at 570 nm using the 630-nm wavelength as reference.

Viability Evaluation in 3D Cell Cultures
Four cell lines, namely HT-29, MDA-MB-231, SKOV3, and NCI-H23, were cultured at 37 • C and at 5% CO 2 using RPMI-1640, Dulbecco's Modified Eagle's medium (DMEM), McCoy's 5A and RPMI-1640 medium respectively, supplemented with 10% fetal bovine serum. When the cells were about 75% confluent in 75 cm 2 flasks, they were trypsinized with trypsin-EDTA, counted, and plated in 96-well plates with round bottom at a density of 10 3 cells/well in a final volume of 100 µL. After 7 days of incubation in a 5% CO 2 humidified incubator at 37 • C, the plates were treated with the same solutions/complexes of CAI presented above. Each assay was done at least in triplicate. After 24 h of incubation, cells are treated with 10% WST-8 in media. The plate was stored in the incubator for 4 h, and the plate was subsequently read for absorbance at 450 nm with a reference at 650 nm.

CA IX Profiling Using Western Blotting
The HT-29, MDA-MB231, SKOV-3, and NCI-H23 cell lines were screened for the expression of carbonic anhydrase IX in 2D and 3D cell culture under normoxic and hypoxic conditions. Cells were cultured in 60-mm petri dishes at a density of 10 6 cells under normoxic and hypoxic conditions. The hypoxic conditions were induced as described above by placing the cells into a hypoxic chamber purged with hypoxia gas mixture containing 1% oxygen for 24 h. The chamber was placed in a 37 • C incubator. For normoxic conditions, cells were kept in a regular incubator for 24 h. The next day, all cells were harvested, pelleted, and washed with phosphate buffer saline (PBS). Cells were then lysed with radioimmunoprecipitation assay (RIPA) buffer containing protease inhibitors, and lysates were collected and stored at −20 • C. Total protein concentration was determined using the bicinchoninic acid method (BCA). The cells were also cultured in rounded bottom 96-well plates at a density of 1000 cells/well and were placed in the incubator for 1-2 week to grow into tumor spheroids. After the spheroids were fully developed, they were harvested and lysed as mentioned previously. Western Blot analysis was performed as described [52]. Briefly, cell lysate samples were loaded onto 10% SDS-PAGE precast gels (Bio-rad, 456-8034), and separation was done at 150 V for 2 h. Gels were transblotted onto nitrocellulose membranes for 30 min, and membranes were blocked using the blocking buffer followed by incubation with a specific primary mouse monoclonal antibody for CA IX (M75, Bioscience Slovakia) at 4 • C overnight. After washing, the membranes were incubated with anti-mouse IgG IRDye800CW secondary antibody (Rockland) for 1 h at room temperature. Finally, membranes were detected using an Odyssey image system (Li-Cor, Lincoln, NE, USA). Beta actin was used as a control and was detected using a mouse monoclonal beta actin antibody (Genetex) binding followed by detection with the same IgG IRDye800CW secondary antibody.

Conclusions
A series of bifunctional PEGylated CAIs 28-31 congeners of previously reported representative 23 was generated following our established synthetic strategy [52]. The series allowed us to test whether the PEG linker separating the two individual warheads elicit significant effects on the biological properties of this type of CAI. DSC data allowed us to conclude that, starting from a 3.4 KDa linker length, the two warheads are spaced enough to have minimum influence on each other and on the PEG backbone. This linker length was found relevant for the ability of CAIs to kill cancer cells in 2D and 3D (tumor spheroids) that express CA IX. Thus, PEGylated CAIs with short linkers were more efficient in killing these cancer cells in vitro than their congeners with long PEG linkers, confirming a strong impact of cooperativity towards CA IX inhibition with these dimeric CAIs. Tumor cell killing was found to be amplified under hypoxia when the expressions of CA IX and CA XII are strongly induced. Also, the polymeric CAIs did not affect the cell viability of CA IX-negative NCI-H23 cells, thus confirming CA IX-mediated cell killing of these inhibitors.

Data Availability Statement:
The data presented in this study are available in the article. Additional data are available from the corresponding author upon request.