Obtaining control of cell surface functionalizations via Pre-targeting and Supramolecular host guest interactions

The use of mammalian cells for therapeutic applications is finding its way into modern medicine. However, modification or “training” of cells to make them suitable for a specific application remains complex. By envisioning a chemical toolbox that enables specific, but straight-forward and generic cellular functionalization, we investigated how membrane-receptor (pre)targeting could be combined with supramolecular host-guest interactions based on β-cyclodextrin (CD) and adamantane (Ad). The feasibility of this approach was studied in cells with membranous overexpression of the chemokine receptor 4 (CXCR4). By combining specific targeting of CXCR4, using an adamantane (Ad)-functionalized Ac-TZ14011 peptide (guest; KD = 56 nM), with multivalent host molecules that entailed fluorescent β-CD-Poly(isobutylene-alt-maleic-anhydride)-polymers with different fluorescent colors and number of functionalities, host-guest cell-surface modifications could be studied in detail. A second set of Ad-functionalized entities enabled introduction of additional surface functionalities. In addition, the attraction between CD and Ad could be used to drive cell-cell interactions. Combined we have shown that supramolecular interactions, that are based on specific targeting of an overexpressed membrane-receptor, allow specific and stable, yet reversible, surface functionalization of viable cells and how this approach can be used to influence the interaction between cells and their surroundings.

propylamine·HBr (0.87 g, 4 mmol) in 10 mL 1,2-dichlorobenzene was stirred for 30 min at 110 o C, followed by 10 min at 150 o C. The resulting purple precipitate was collected and dispersed in 15 mL MeOH. Di-tert-butyldicarbonate (1.7 g, 8 mmol) and DIPEA (1.4 mL, 8 mmol) were added and the reaction mixture was refluxed for 30 minutes. The mixture was concentrated in vacuo and purified by column chromatography (MeOH:CH 2 Cl 2 1:3), yielding the product as a pink solid (81 mg).
DIPEA (181 µL, 1.04 mmol) and acetic anhydride (59 µL, 0.63 mmol) were added and the mixture was stirred for 2 h at room temperature. The mixture was concentrated under vacuum yielding a colorless oil. The oil was re-dissolved in 20 mL EtOH together with indole-COOH (127 mg, 0.46 mmol), indole-Phth (197 mg, 0.46 mmol) and pyridine (169 μL, 2.10 mmol). The solution was refluxed for 3 h and stirred overnight at 60 °C. Acetic anhydride (60 μL) was added, and the reaction mixture turned pink. After refluxing for 4 h, the crude product was concentrated in vacuo and purified by column chromatography (eluens: MeOH). The product fractions were combined, concentrated and lyophilized to give a pink solid of impure compound Cy3-Phth-COOH. A portion of Cy3-Phth-COOH (100 mg, 0.16 mmol) was further purified by preparative HPLC. After lyophilization of the product fractions, Cy3-Phth-COOH was deprotected by adding 2 mL of CH 3 NH 2 (33% in EtOH). The solution was stirred for 5 h, after which the reaction was concentrated in vacuo to give Cy3-Amine-COOH as a pink solid.
Subsequently the compound was purified by preparative HPLC. Fraction containing product was collected and lyophilized and gave the pure product as a pink solid (4 mg, 7.99 µmol, 5% yield). For detailed description of the polymers synthesis, see main manuscript

Analysis of the CD-polymers
The grafting efficiency of β-CD and the fluorophores was determined by a combination of 1 H-NMR and UV/Vis absorption measurements. The grafting of the β-CD was determined by 1 H-NMR, by integrating the polymer peaks at 1.38 -1.00 ppm (both methyl and CH 2 moieties) and the β-CD peaks at 5.1 ppm (anomeric carbon CH) and 4.00 -3.50 ppm (all other β-CD protons).
The integral of the peaks corresponding to the polymer was then set at 8 (Suplementary Fig.   S10B). The obtained integral for the β-CD peaks at 5.1 ppm and 4.00 -3.50 ppm were divided by

Comparison of multivalent and monovalent functionalization
To study the differences in binding of multivalent-and monovalent β-CD compounds to Ac-TZ14011-Ad functionalized cells, adhering MDAMB231 X4 cells were functionalized with either The replacement between Cy5 0.5 CD 10 PIBMA 39 and Cy3 1.5 CD 72 PIBMA 389 , was performed according the same procedure as described for the competition experiment followed over time.
Only now the excess of binding polymer was washed away with PBS (2 x 1 mL) before imaging.
As a control, besides the respective polymers, also only medium was added. The change in fluorescence was followed for 18 minutes, while taking images each minute ( Supplementary   Fig. S18 and S19).
The replacement of the non-host containing Cy5 0.4 PIBMA 39 polymer by  Fig. S18 and S19).

Functionalization of human stem cells
Human fetal heart stem cells, with CXCR4 expression, (17 weeks after gestation) were grown in M199 -/-on gelatin-coated glass-bottom dishes (100,000 cells per dish). These cells were functionalized with Ac-TZ14011-Ad and Cy3 1.5 CD 72 PIBMA 389 according described procedure (see 'functionalization of cells', main manuscript). Subsequently, they were carefully washed with colorless DMEM (2 x 1 mL) and analyzed by confocal microscopy (Supplementary Fig. S22).
Hereafter cells were divided into aliquots (300,000 cells per tube), centrifuged for three minutes (3000

Determination of the receptor affinity of Ac-TZ14011-Ad
The affinity (K D ) of Ac-TZ14011-Ad was calculated from flow cytometry measurements, using an earlier described procedure. 8 In short: Different concentrations of Ac-TZ14011-Ad, ranging between 0.5 -15,000 nM in 120 μL PBS, were added to MDAMB231 X4 cells in the presence of Ac-TZ14011-MSAP (250 nM), a compound with well-defined receptor affinity. 8 After one hour of incubation on ice, the cells were washed two times with PBS (centrifuged 3 min, 3000 x g, 4 °C), and resuspended in 300 µL PBS. The fluorescence of the reference compound was measured as described in the flow cytometry section. All experiments were performed in duplicate (n = 2). The mean fluorescence was normalized and fitted with equations in the GraphPad Prism 6 software ( Supplementary Fig. S11). The K D values were calculated using the "Binding-Competitive, One site-Fit Ki" nonlinear regression equation (Equation (3) and (4) were counted for radioactivity in a dose-calibrator or gamma counter to assess the amount of cellular-bound 99m Tc-Cy n -CD x -PIBMA y activity. Data was expressed as the mean % (±SD, n = 6) of the total amount of 99m Tc-Cy n -CD x -PIBMA y activity added to the cells (Supplementary Fig. S16).

Cell viability assay
To determine the effect of polymer functionalization on the cell viability, after cell functionalization a MTT test was performed according to a described procedure. 9

Synthesis of Cy5-CD, Cy5-Ad and Cy5Ad 2
To compare cell functionalization with multivalent β-CD groups (Cy5 0.5 CD 10 PIBMA 39 , Cy3 1.5 CD 72 PIBMA 389 ) to monovalent β-CD, a Cy5 fluorophore was conjugated to β-CD via amide bond formation to give Cy5-CD (52% yield). By the same synthesis strategy Cy5-Ad was synthesized (57% yield), as mode to investigate if polymer coated cells could further be functionalized via de CD-Ad host-guest interaction. To strengthen this interaction also a Cy5-Ad 2 was synthesized (2.6% yield). Both Ad molecules were connected to the Cy5 and each other via a short spacer consisting of 2 beta alanines, 7 again functionalization could be established via multiple amide bond formations.

Functionalization: mono vs multivalent β-CD hosts
Comparing the fluorescence signal intensities after functionalization with Cy5-CD (monovalent host) or Cy5 0.5 CD 10 PIBMA 39 (multivalent host), reveals substantial more binding to Ad functionalized cell surfaces by Cy5 0.5 CD 10 PIBMA 39 (supplementary Fig. S14 and S15). As a control the same functionalization was also performed with Cy5 0.4 PIBMA (Supplementary Fig   S7). While Cy5 0.4 PIBMA 39 binds slightly more to the cell surface compared to Cy5-CD, it are the multiple CD groups on Cy5 0.5 CD 10 PIBMA 39 that considerably increases the amount of binding.

Radiolabeling of the polymers
Cy5 0.5 CD 10 PIBMA 39 , and Cy3 1.5 CD 72 PIBMA 389 were radiolabeled with technetium-99m according to a modification of a previously described labeling technique (described above). 11 We propose that the carboxylates, formed after hydrolysis, and possibly also the side chains of β-CD could function as chelate for the technetium-99m ions after reduction with SnCl 2 . Since no significant release of radioactivity from the polymer was observed until 24 hours after radiolabeling, the interaction was stable under the in vitro conditions here applied.

Exchange and competition between CD n PIBMA m polymers
To  Fig. S18D). To validate these findings the experiment was also performed in reversed order (Cy3 1.5 CD 72 PIBMA 389 as initial binder and Cy5 0.4 PIBMA 39 as competitor ( Supplementary Fig. S19D)). Here initial non-specific binding of Cy5 0.4 PIBMA 39 occurred, but no further increase in binding could be observed over time. The  Fig. S19C). This data provides a further indication that Cy3 1.5 CD 72 PIBMA 389 is the better binder of the two.
Combined, these data suggest that the polymer functionalization is stable when no competition occurs, but is reversible under competitive conditions. Furthermore, the data confirms that there is a different mode of action for polymers that contain β-CD vs. those that don't.