# Response Surface Methodology to Efficiently Optimize Intracellular Delivery by Photoporation

^{1}

^{2}

^{*}

^{†}

## Abstract

**:**

## 1. Introduction

## 2. Results

#### 2.1. Synthesis of Polydopamine Nanoparticles

^{2}, 1.12 J/cm

^{2}and 0.95 J/cm

^{2}for 500 nm, 700 nm and 858 nm PDNPs, respectively.

#### 2.2. Optimization of Delivery Yield by Response Surface Methodology

#### 2.2.1. Model Fitting and Statistical Assessment

^{2}) and quadratic fluence term (Fluence

^{2}) to be weakly significant in two of the three replicates, providing no convincing proof of a consistently significant contribution of these variables. Both designs thus led to models that suggest a considerable dependency of the delivery yield on the PDNP size, whereas the PDNP concentration and laser pulse fluence only seem to have relatively minor effects on the investigated response within the defined ranges, with the BBD-based models generally estimating them to be less of importance than the models based on the CCC design.

^{2}) and adjusted R

^{2}(adj. R

^{2}). Higher values of statistics, such as the root mean squared error (RMSE) and mean absolute error (MAE) for the CCC design-based models, also demonstrate a less adequate fit relative to the BBD-based models. Preliminary metrics of the models’ predictive powers, also based on the fit itself, again suggest the BBD result in better fitting models, as reflected by their seemingly higher predicted R

^{2}(pred. R

^{2}) and lower predicted residual sum of squares (PRESS) values. However, relatively small pred. R

^{2}values in contrast to both multiple R

^{2}and adj. R

^{2}, for all cases, signifies potentially tenuous reliability of the models to make predictions in the design space but can only be confirmed by testing the fit on data that was not used to fit the model.

#### 2.2.2. Analysis of the Response Surface

#### 2.2.3. Design Adjustment for the BDD

^{2}and other metrics match those of the original BBD-based models. Nonetheless, the PDNP size was again the only consistent significant estimate in its linear and quadratic terms, with other terms being more variable in their significance. Breaches of the underlying assumptions were not detected.

#### 2.2.4. Model Validation

^{2}, RMSE and MAE) indicating a clear decrease in model performance when evaluating their fit to the confirmation run data (Table S12), these accuracies suggest that the model output of each of the used designs is capable of predicting the yield within the design space with sufficient reliability. No discernable differences were observed between the CCC design-based and BBD-based models–original or extended–regarding their predictive power.

^{8}PDNPs/mL—resulted in considerably higher, and even close to optimal, yields than when suboptimal parameter value combinations—e.g., 300 nm or 700 nm PDNPs, regardless of concentration and fluence—were used. Thus, the models’ identified optimum is successfully corroborated. Further support for this assertion was found in run four for all designs (Table 4, Tables S10 and S11), with a quasi-optimal factor level combination identified by the models, as they resulted in the practically highest obtained delivery yields. Moreover, it became evident that the PDNP size was the main contributor to the delivery yield, whereas the PDNP concentration and laser pulse fluence only demonstrated comparatively minor effects within the explored design space, exactly as predicted.

## 3. Discussion

^{2}values, whether multiple or adjusted and lower values for the RMSE and MAE favor the choice for the BBD in this specific area [62,75,76,77].

^{2}metric for the models resulting from the two designs revealed a severe decrease (>0.20) in comparison to the multiple and adj. R

^{2}, suggesting the low predictive power of the models [78]. Confirmation runs corroborated this finding by also establishing a true sizeable reduction in the multiple R

^{2}using their data points—i.e., a measure of fit between the originally fitted curve and newly collected data points. Possible reasons for this are the presence of irrelevant model terms (i.e., overfitting) [78], model biases due to factor level errors [79] and the possibility that a second-order model does not adequately describe the relationship between the response and factors [80]. Broad design regions—e.g., by ±α-values in the CCC design—can also contribute to this issue [81]. Interestingly, the similarity of the multiple R

^{2}, RMSE and MAE for the confirmatory run replicates of the two designs indicated that they had equal predictive power based on statistical metrics, even though the BBD-based models had a considerable advantage in the first adequacy check. It must, however, be considered that some of the confirmatory runs for the BBD-based models were located outside its design space, thus requiring extrapolation, which is inherently more unreliable [82]. Despite the suggested lackluster predictive capabilities of both designs’ models by several calculated metrics, the high obtained accuracies based on experimental validation data points provided a first indication that the models can indeed reliably predict experimental regions of interest to some extent [83]. No difference between the outputs of the designs could be made in this aspect.

^{8}NPs/mL as optimal instead of ca. 2.5 × 10

^{8}NPs/mL by the CCC design-based models. In practice, however, these are rather small differences that are likely close to optimal. A comparison of the delivery yields of predicted near-optimal parameter value combinations to predicted suboptimal combinations did indeed confirm that optimum identification was effective for the models resulting from both designs. This further strengthened the preceding model validation based on confirmation runs. Considering that both designs resulted in models that were equally successful in identifying an optimum and the aforementioned effective practical validation, it is shown that (better) general statistical metrics do not necessarily lead to better/proper predictive capabilities in practice, possibly warranting some practical liberties. Additionally, the similarity of the discerned optimal conditions between the different design replicates proved that only one designed experiment is sufficient, confirming one of the key concepts in RSM. Lastly, a note on the design selection has to be made, however. Even though both designs were successfully used for optimization purposes in this study, the BBD should only be selected if the experiment permits little flexibility, as a CCC design has better predictive capacities across the entire design space [62]. Carefully made considerations of the experimental design remain crucial.

^{2}, for which the balance of heating vs. VNB formation is fairly similar per PDNP size (Figure 3). It is, therefore, understandable that it had little influence on the final delivery yield. For a given fluence, however, there is a big difference in the predominant photothermal mechanism depending on the PDNP size. For the smaller PDNPs, heating is more prevalent, while VNB formation will be more likely for the larger PDNP sizes. Within the tested concentration range, permeabilization by VNBs appears to be more beneficial than by heating, an observation that was made before for AuNPs [19]. The fact that 500 nm PDNPs come out as more beneficial than larger ones, which also form VNBs, is likely related to the fact that they may give rise to larger and more powerful VNBs, which can lead to more damage to cells [20]. Indeed, delivery yield is the balance between delivery efficiency and cell viability. Within the tested concentration range, PDNPs larger than 500 nm inflict damage to cells too extensively (Figure S6), which negatively impacts delivery yield. It cannot be excluded that higher (still sub-optimal) delivery yields are possible for those large PDNPs if the concentration range is expanded to include lower concentrations so as to reduce toxic effects. Similarly, for higher PDNP concentrations, it may be that better, but perhaps still sub-optimal, yields are possible for the smaller PDNPs. This remains to be explored in future research.

## 4. Materials and Methods

#### 4.1. Synthesis of Polydopamine Nanoparticles and Functionalization with Bovine Serum Albumin

#### 4.2. Physicochemical and Morphological Characterization

#### 4.3. VNB-Threshold

^{9}NPs/mL in double-distilled water (ddH

_{2}O), present in a 50 mm γ-irradiated glass bottom dish (MatTek Corporation, Ashland, MA, USA), with ca. 3 ns pulsed 532 nm laser light (Cobolt TorTM Series, Cobolt AB, Solna, Sweden). Lasers pulses were generated on demand using a 25 MHz pulse generator (TGP3121, Aim-TTi, Huntingdon, UK), with control over the pulse energy being provided by an adjustable DC power supply (HQ Power PS23023, Velleman Group, Gavere, Belgium). Energies were registered using an Ophir Starlite energy meter (MKS Instruments, Andover, MA, USA). The VNBs were visualized using dark field microscopy, where the increased scatter of VNBs resulted in bright white spots [19]. Short movies of this phenomenon were captured using a cMOS camera (Blackfly S GigE-Mono, FLIR, Wilsonville, OR, USA) and screen recording software, which allowed the counting of individual VNBs. The number of generated VNBs was determined in the irradiated area as a function of the applied laser fluence. A Boltzmann sigmoid curve was fitted to the data normalized against the maximal number of counted VNBs in GraphPad Prism version 8.0.0 (GraphPad Software, San Diego, CA, USA), allowing quantification of the VNB-threshold as the laser pulse fluence at which 90% of the irradiated particles generate a detectable VNB.

#### 4.4. Cell Culture

^{TM}-I (Gibco

^{TM}), which was supplemented with 10% fetal bovine serum (FBS, Biowest, Nuaillé, France) and 100 U/mL Penicillin/Streptomycin (Gibco

^{TM}). Every 2 or 3 days, the cells reached confluency and were subsequently passaged and kept in a humidified incubator (37 °C, 5% CO

_{2}).

#### 4.5. Photoporation for the Intracellular Delivery of FITC-Dextran 500 kDa

_{2}). Next, the cells were washed once with Dulbecco’s phosphate buffered saline without Ca

^{2+}and Mg

^{2+}[DPBS(-/-), Biowest, Nuaillé, France] [89] after which the PDNPs, of which a series of dilutions was prepared in reduced serum culture medium [Optimized-Minimal Essential Medium (Opti-MEM), Gibco

^{TM}], were added to the wells. This step was immediately followed by 10 s centrifugation in a plate centrifuge (Eppendorf, Hamburg, Germany) until 1300 RCF was reached. Unbound particles were removed from the cells by washing them with DPBS(-/-). 50 µL of Opti-MEM containing 1 mg/mL FITC-dextran 500 kDa (FD500, Sigma-Aldrich, Bornem, Belgium) was then added to the wells, followed by photoporation with an in-house developed optical setup with a nanosecond laser (3 ns pulse duration, 532 nm wavelength) and equipped with a Galvano-scanner to rapidly scan the laser beam across the samples (5–6 s per well). Immediately after laser treatment, cells were washed twice with fresh culture medium to remove excess FD500 and thus prevent its spontaneous uptake by endocytic processes. Finally, the cells were prepared for flow cytometry analysis of the delivery efficiency.

#### 4.6. Analysis of Intracellular Delivery Efficiency by Flow Cytometry

^{TM}) for 5 min on an orbital shaker (120 rpm) and then adding 110 µL of flow buffer [DPBS(-/-), 1% Bovine Serum Albumin, 0.1% Sodium Azide] containing 0.5 × 10

^{−6}M of a cyanine dye monomer TO-PRO3 iodide (Invitrogen, Belgium) as cell viability dye. The samples were then measured with a CytoFLEX flow cytometry (Beckman Coulter, Krefeld, Germany). For every sample, 10,000 living cells were gated for the evaluation of the FD500 delivery in the FITC-channel by excluding TO-PRO3-positive cells in the singlet population. FlowJo™ software (Treestar Inc., Ashland, OR, USA) was used for data analysis.

#### 4.7. Cell Viability Analysis

_{2}) for 2 to 4 h, after which the viability was assessed using the CellTiter Glo

^{®}luminescent cell viability assay (Promega, Belgium). In this assay, the number of viable cells is determined by quantification of the ATP present after cell lysis, as an indicator of the metabolic activity of cells. As recommended by the manufacturer, the old culture medium was removed, and 100 µL of fresh culture medium was added, which was supplemented by an equal volume of CellTiter Glo

^{®}reagent and shaken on an orbital shaker (120 rpm) for 10 min at room temperature. Next, the cell lysates were transferred to an opaque white 96-well plate (Greiner Bio-One, Belgium), and the luminescent signal was measured using a GloMax

^{®}microplate reader (Promega, Belgium). Cell viability was calculated relative to the average of a technical triplicate of non-treated controls.

#### 4.8. Response Surface Methodology

#### 4.8.1. Experimental Design

^{®}assay. Both the five-factor level Circumscribed Central Composite (CCC) design and three-factor level Box-Behnken Design (BBD) (Table S13) were generated in R (v4.2.1) using the rsm package (v2.10.3) with four center-point runs (i.e., the center of design space) [67]. Randomization of the run order was performed to prevent introducing bias in the data by a time-effect. Table 5 displays the parameter values included in the design, which were selected based on prior knowledge of typical photoporation parameters, and the corresponding coded values as used in the RSM analysis, which were defined as

#### 4.8.2. Model Fitting and Statistical Assessment

^{2}), adjusted coefficients of determination (adj. R

^{2}), predicted R

^{2}(pred. R

^{2}), predicted residual sum of squares (PRESS), root mean squared error (RMSE) and mean absolute error (MAE) [62,75,76,92]. The normal distribution of the residuals was checked by calculating the residual mean, plotting the relation between the theoretical normal quantiles and the ordered normalized residuals (Q-Q plot) and the Shapiro-Wilk test of normality [62,93]. Homoscedasticity was tested using the Bruesch-Pagan test [94]. Lastly, the introduction of bias into the model by a time dependency was checked by the Durbin-Watson test and plotting the residuals as a function of the run order [62,95]. For all statistical tests, the null hypothesis was rejected when the p-value < 0.05.

#### 4.8.3. Canonical Analysis

#### 4.8.4. Model Validation

^{2}, RMSE and MAE on the validation data set.

## 5. Conclusions

## Supplementary Materials

## Author Contributions

## Funding

## Institutional Review Board Statement

## Informed Consent Statement

## Data Availability Statement

## Conflicts of Interest

## References

- Pei, D. How do biomolecules cross the cell membrane? Acc. Chem. Res.
**2022**, 55, 309–318. [Google Scholar] [CrossRef] [PubMed] - Morshedi Rad, D.; Alsadat Rad, M.; Razavi Bazaz, S.; Kashaninejad, N.; Jin, D.; Ebrahimi Warkiani, M. A Comprehensive review on intracellular delivery. Adv. Mater.
**2021**, 33, 2005363. [Google Scholar] [CrossRef] [PubMed] - Yang, N.J.; Hinner, M.J. Getting Across the Cell Membrane: An Overview for Small Molecules, Peptides, and Proteins. In Methods in Molecular Biology (Clifton N.J.); Humana Press: New York, NY, USA, 2015; Volume 1266, pp. 29–53. [Google Scholar]
- Butt, M.; Zaman, M.; Ahmad, A.; Khan, R.; Mallhi, T.; Hasan, M.; Khan, Y.; Hafeez, S.; Massoud, E.; Rahman, M.; et al. Appraisal for the potential of viral and nonviral vectors in gene therapy: A review. Genes
**2022**, 13, 1370. [Google Scholar] [CrossRef] [PubMed] - Wang, X.; Li, C.; Wang, Y.; Chen, H.; Zhang, X.; Luo, C.; Zhou, W.; Li, L.; Teng, L.; Yu, H.; et al. Smart drug delivery systems for precise cancer therapy. Acta Pharm. Sin. B
**2022**, 12, 4098–4121. [Google Scholar] [CrossRef] - Mitchell, M.J.; Billingsley, M.M.; Haley, R.M.; Wechsler, M.E.; Peppas, N.A.; Langer, R. engineering precision nanoparticles for drug delivery. Nat. Rev. Drug Discov.
**2021**, 20, 101–124. [Google Scholar] [CrossRef] - Yin, H.; Kanasty, R.L.; Eltoukhy, A.A.; Vegas, A.J.; Dorkin, J.R.; Anderson, D.G. Non-Viral vectors for gene-based therapy. Nat. Rev. Genet.
**2014**, 15, 541–555. [Google Scholar] [CrossRef] - Stewart, M.P.; Langer, R.; Jensen, K.F. intracellular delivery by membrane disruption: Mechanisms, strategies, and concepts. Chem. Rev.
**2018**, 118, 7409–7531. [Google Scholar] [CrossRef] - Shi, J.; Ma, Y.; Zhu, J.; Chen, Y.; Sun, Y.; Yao, Y.; Yang, Z.; Xie, J. A review on electroporation-based intracellular delivery. Molecules
**2018**, 23, 3044. [Google Scholar] [CrossRef] - Kooiman, K.; Roovers, S.; Langeveld, S.A.G.; Kleven, R.T.; Dewitte, H.; O’Reilly, M.A.; Escoffre, J.-M.; Bouakaz, A.; Verweij, M.D.; Hynynen, K.; et al. Ultrasound-responsive cavitation nuclei for therapy and drug delivery. Ultrasound Med. Biol.
**2020**, 46, 1296–1325. [Google Scholar] [CrossRef] - Bi, Q.; Song, X.; Hu, A.; Luo, T.; Jin, R.; Ai, H.; Nie, Y. Magnetofection: Magic Magnetic Nanoparticles for Efficient Gene Delivery. Chinese Chem. Lett.
**2020**, 31, 3041–3046. [Google Scholar] [CrossRef] - Chakrabarty, P.; Gupta, P.; Illath, K.; Kar, S.; Nagai, M.; Tseng, F.-G.; Santra, T.S. Microfluidic mechanoporation for cellular delivery and analysis. Mater. Today Bio
**2022**, 13, 100193. [Google Scholar] [CrossRef] [PubMed] - Ramon, J.; Xiong, R.; De Smedt, S.C.; Raemdonck, K.; Braeckmans, K. Vapor nanobubble-mediated photoporation constitutes a versatile intracellular delivery technology. Curr. Opin. Colloid Interface Sci.
**2021**, 54, 101453. [Google Scholar] [CrossRef] - Xiong, R.; Samal, S.K.; Demeester, J.; Skirtach, A.G.; De Smedt, S.C.; Braeckmans, K. Laser-assisted photoporation: Fundamentals, technological advances and applications. Adv. Phys. X
**2016**, 1, 596–620. [Google Scholar] [CrossRef] - Wang, D.; Cheng, P.; Quan, X. Photothermal nanobubble nucleation on a plasmonic nanoparticle: A 3D lattice boltzmann simulation. Int. J. Heat Mass Transf.
**2019**, 140, 786–797. [Google Scholar] [CrossRef] - Lombard, J.; Lam, J.; Detcheverry, F.; Biben, T.; Merabia, S. Strong and fast rising pressure waves emitted by plasmonic vapor nanobubbles. Phys. Rev. Res.
**2021**, 3, 023231. [Google Scholar] [CrossRef] - Maheshwari, S.; van der Hoef, M.; Prosperetti, A.; Lohse, D. Dynamics of formation of a vapor nanobubble around a heated nanoparticle. J. Phys. Chem. C
**2018**, 122, 20571–20580. [Google Scholar] [CrossRef] - Lukianova-Hleb, E.; Hu, Y.; Latterini, L.; Tarpani, L.; Lee, S.; Drezek, R.A.; Hafner, J.H.; Lapotko, D.O. Plasmonic nanobubbles as transient vapor nanobubbles generated around plasmonic nanoparticles. ACS Nano
**2010**, 4, 2109–2123. [Google Scholar] [CrossRef] - Xiong, R.; Raemdonck, K.; Peynshaert, K.; Lentacker, I.; De Cock, I.; Demeester, J.; De Smedt, S.C.; Skirtach, A.G.; Braeckmans, K. Comparison of gold nanoparticle mediated photoporation: Vapor nanobubbles outperform direct heating for delivering macromolecules in live cells. ACS Nano
**2014**, 8, 6288–6296. [Google Scholar] [CrossRef] [PubMed] - Kumar, S.; Chong, C.; Voorhees, T.; Thadhani, N.N.; Prausnitz, M.R. Relationship between bio-effects and energy transduction during nanoparticle-mediated photoporation. J. Appl. Phys.
**2020**, 128, 173101. [Google Scholar] [CrossRef] - Xiong, R.; Hua, D.; Van Hoeck, J.; Berdecka, D.; Léger, L.; De Munter, S.; Fraire, J.C.; Raes, L.; Harizaj, A.; Sauvage, F.; et al. Photothermal nanofibres enable safe engineering of Therapeutic cells. Nat. Nanotechnol.
**2021**, 16, 1281–1291. [Google Scholar] [CrossRef] - Raes, L.; Pille, M.; Harizaj, A.; Goetgeluk, G.; Van Hoeck, J.; Stremersch, S.; Fraire, J.C.; Brans, T.; de Jong, O.G.; Maas-Bakker, R.; et al. Cas9 RNP Transfection by Vapor Nanobubble Photoporation for Ex Vivo Cell Engineering. Mol. Ther.-Nucleic Acids
**2021**, 25, 696–707. [Google Scholar] [CrossRef] [PubMed] - Harizaj, A.; Descamps, B.; Mangodt, C.; Stremersch, S.; Stoppa, A.; Balcaen, L.; Brans, T.; De Rooster, H.; Devriendt, N.; Fraire, J.C.; et al. Cytosolic delivery of gadolinium via photoporation enables improved in vivo magnetic resonance imaging of cancer cells. Biomater. Sci.
**2021**, 9, 4005–4018. [Google Scholar] [CrossRef] - Harizaj, A.; Van Hauwermeiren, F.; Stremersch, S.; De Rycke, R.; De Keersmaecker, H.; Brans, T.; Fraire, J.C.; Grauwen, K.; De Smedt, S.C.; Lentacker, I.; et al. Nanoparticle-sensitized photoporation enables inflammasome activation studies in targeted single cells. Nanoscale
**2021**, 13, 6592–6604. [Google Scholar] [CrossRef] [PubMed] - Raes, L.; Stremersch, S.; Fraire, J.C.; Brans, T.; Goetgeluk, G.; De Munter, S.; Van Hoecke, L.; Verbeke, R.; Van Hoeck, J.; Xiong, R.; et al. Intracellular delivery of MRNA in adherent and suspension cells by vapor nanobubble photoporation. Nano-Micro Lett.
**2020**, 12, 1–17. [Google Scholar] [CrossRef] [PubMed] - Raes, L.; Van Hecke, C.; Michiels, J.; Stremersch, S.; Fraire, J.C.; Brans, T.; Xiong, R.; De Smedt, S.; Vandekerckhove, L.; Raemdonck, K.; et al. Gold nanoparticle-mediated photoporation enables delivery of macromolecules over a wide range of molecular weights in human CD4+ T cells. Crystals
**2019**, 9, 411. [Google Scholar] [CrossRef] - Wang, L.; Wu, J.; Hu, Y.; Hu, C.; Pan, Y.; Yu, Q.; Chen, H. Using porous magnetic iron oxide nanomaterials as a facile photoporation nanoplatform for macromolecular delivery. J. Mater. Chem. B
**2018**, 6, 4427–4436. [Google Scholar] [CrossRef] [PubMed] - Wayteck, L.; Xiong, R.; Braeckmans, K.; De Smedt, S.C.; Raemdonck, K. Comparing Photoporation and nucleofection for delivery of small interfering RNA to Cytotoxic T cells. J. Control. Release
**2017**, 267, 154–162. [Google Scholar] [CrossRef] - Kumar, S.; Li, A.; Thadhani, N.N.; Prausnitz, M.R. Optimization of Intracellular macromolecule delivery by nanoparticle-mediated photoporation. Nanomed. Nanotechnol. Biol. Med.
**2021**, 37, 102431. [Google Scholar] [CrossRef] - Liu, J.; Li, C.; Brans, T.; Harizaj, A.; Van de Steene, S.; De Beer, T.; De Smedt, S.; Szunerits, S.; Boukherroub, R.; Xiong, R.; et al. Surface functionalization with polyethylene glycol and polyethyleneimine improves the performance of graphene-based materials for safe and efficient intracellular delivery by laser-induced photoporation. Int. J. Mol. Sci.
**2020**, 21, 1540. [Google Scholar] [CrossRef] - Teirlinck, E.; Barras, A.; Liu, J.; Fraire, J.C.; Lajunen, T.; Xiong, R.; Forier, K.; Li, C.; Urtti, A.; Boukherroub, R.; et al. Exploring light-sensitive nanocarriers for simultaneous triggered antibiotic release and disruption of biofilms upon generation of laser-induced vapor nanobubbles. Pharmaceutics
**2019**, 11, 201. [Google Scholar] [CrossRef] [Green Version] - Liu, J.; Xiong, R.; Brans, T.; Lippens, S.; Parthoens, E.; Zanacchi, F.C.; Magrassi, R.; Singh, S.K.; Kurungot, S.; Szunerits, S.; et al. Repeated photoporation with graphene quantum dots enables homogeneous labeling of live cells with extrinsic markers for fluorescence microscopy. Light Sci. Appl.
**2018**, 7, 47. [Google Scholar] [CrossRef] [PubMed] - Holguin, S.Y.; Thadhani, N.N.; Prausnitz, M.R. Effect of laser fluence, nanoparticle concentration and total energy input per cell on photoporation of cells. Nanomed. Nanotechnol. Biol. Med.
**2018**, 14, 1667–1677. [Google Scholar] [CrossRef] [PubMed] - Xiong, R.; Joris, F.; Liang, S.; De Rycke, R.; Lippens, S.; Demeester, J.; Skirtach, A.; Raemdonck, K.; Himmelreich, U.; De Smedt, S.C.; et al. Cytosolic delivery of nanolabels prevents their asymmetric inheritance and enables extended quantitative in vivo cell imaging. Nano Lett.
**2016**, 16, 5975–5986. [Google Scholar] [CrossRef] [PubMed] - Sengupta, A.; Kelly, S.C.; Dwivedi, N.; Thadhani, N.; Prausnitz, M.R. Efficient intracellular delivery of molecules with high cell viability using nanosecond-pulsed laser-activated carbon nanoparticles. ACS Nano
**2014**, 8, 2889–2899. [Google Scholar] [CrossRef] - Chakravarty, P.; Qian, W.; El-Sayed, M.A.; Prausnitz, M.R. Delivery of molecules into cells using carbon nanoparticles activated by femtosecond laser pulses. Nat. Nanotechnol.
**2010**, 5, 607–611. [Google Scholar] [CrossRef] - Di Ianni, E.; Jacobsen, N.R.; Vogel, U.B.; Møller, P. Systematic review on primary and secondary genotoxicity of carbon black nanoparticles in mammalian cells and animals. Mutat. Res.
**2022**, 790, 108441. [Google Scholar] [CrossRef] - Ghulam, A.N.; dos Santos, O.A.L.; Hazeem, L.; Pizzorno Backx, B.; Bououdina, M.; Bellucci, S. Graphene oxide (GO) materials—Applications and toxicity on living organisms and environment. J. Funct. Biomater.
**2022**, 13, 77. [Google Scholar] [CrossRef] - Sani, A.; Cao, C.; Cui, D. Toxicity of Gold Nanoparticles (AuNPs): A review. Biochem. Biophys. Rep.
**2021**, 26, 100991. [Google Scholar] [CrossRef] - Mohammadpour, R.; Dobrovolskaia, M.A.; Cheney, D.L.; Greish, K.F.; Ghandehari, H. Subchronic and chronic toxicity evaluation of inorganic nanoparticles for delivery applications. Adv. Drug Deliv. Rev.
**2019**, 144, 112–132. [Google Scholar] [CrossRef] - Ou, L.; Song, B.; Liang, H.; Liu, J.; Feng, X.; Deng, B.; Sun, T.; Shao, L. Toxicity of graphene-family nanoparticles: A general review of the origins and mechanisms. Part. Fibre Toxicol.
**2016**, 13, 57. [Google Scholar] [CrossRef] [Green Version] - Pan, Y.; Neuss, S.; Leifert, A.; Fischler, M.; Wen, F.; Simon, U.; Schmid, G.; Brandau, W.; Jahnen-Dechent, W. Size-dependent cytotoxicity of gold nanoparticles. Small
**2007**, 3, 1941–1949. [Google Scholar] [CrossRef] [PubMed] - Tsoli, M.; Kuhn, H.; Brandau, W.; Esche, H.; Schmid, G. Cellular uptake and toxicity of Au55 clusters. Small
**2005**, 1, 841–844. [Google Scholar] [CrossRef] [PubMed] - Poinard, B.; Lam, S.A.E.; Neoh, K.G.; Kah, J.C.Y. Mucopenetration and biocompatibility of polydopamine surfaces for delivery in an ex vivo porcine bladder. J. Control. Release
**2019**, 300, 161–173. [Google Scholar] [CrossRef] [PubMed] - Poinard, B.; Zhan, S.; Neo, Y.; Li, E.; Yeo, L.; Peng, H.; Heng, S.; Gee Neoh, K.; Chen, J.; Kah, Y. Polydopamine nanoparticles enhance drug release for combined photodynamic and photothermal therapy. ACS Appl. Mater. Interfaces
**2018**, 10, 21125–21136. [Google Scholar] [CrossRef] [PubMed] - Liu, Y.; Ai, K.; Lu, L. Polydopamine and its derivative materials: Synthesis and promising applications in energy, environmental, and biomedical fields. Chem. Rev.
**2014**, 114, 5057–5115. [Google Scholar] [CrossRef] - Postma, A.; Yan, Y.; Wang, Y.; Zelikin, A.N.; Tjipto, E.; Caruso, F. Self-polymerization of dopamine as a versatile and robust technique to prepare polymer capsules. Chem. Mater.
**2009**, 21, 3042–3044. [Google Scholar] [CrossRef] - Zmerli, I.; Michel, J.-P.; Makky, A. Bioinspired polydopamine nanoparticles: Synthesis, nanomechanical properties, and efficient pegylation strategy. J. Mater. Chem. B
**2020**, 8, 4489–4504. [Google Scholar] [CrossRef] - Liebscher, J. Chemistry of polydopamine—Scope, variation, and limitation. Eur. J. Org. Chem.
**2019**, 2019, 4976–4994. [Google Scholar] [CrossRef] - Wang, X.; Chen, Z.; Yang, P.; Hu, J.; Wang, Z.; Li, Y. Size Control synthesis of melanin-like polydopamine nanoparticles by tuning radicals. Polym. Chem.
**2019**, 10, 4194–4200. [Google Scholar] [CrossRef] - Ryu, J.H.; Messersmith, P.B.; Lee, H. Polydopamine surface chemistry: A decade of discovery. ACS Appl. Mater. Interfaces
**2018**, 10, 7523–7540. [Google Scholar] [CrossRef] - Ju, K.Y.; Lee, Y.; Lee, S.; Park, S.B.; Lee, J.K. Bioinspired polymerization of dopamine to generate melanin-like nanoparticles having an excellent free-radical-scavenging property. Biomacromolecules
**2011**, 12, 625–632. [Google Scholar] [CrossRef] - Harizaj, A.; Wels, M.; Raes, L.; Stremersch, S.; Goetgeluk, G.; Brans, T.; Vandekerckhove, B.; Sauvage, F.; De Smedt, S.C.; Lentacker, I.; et al. Photoporation with biodegradable polydopamine nanosensitizers enables safe and efficient delivery of MRNA in human T cells. Adv. Funct. Mater.
**2021**, 31, 2102472. [Google Scholar] [CrossRef] - Czitrom, V. One-factor-at-a-time versus designed experiments. Am. Stat.
**1999**, 53, 126. [Google Scholar] [CrossRef] - Lendrem, D.W.; Lendrem, B.C.; Woods, D.; Rowland-Jones, R.; Burke, M.; Chatfield, M.; Isaacs, J.D.; Owen, M.R. Lost in space: Design of experiments and scientific exploration in a hogarth universe. Drug Discov. Today
**2015**, 20, 1365–1371. [Google Scholar] [CrossRef] - Box, G.E.P.; Wilson, K.B. On the experimental attainment of optimum conditions. J. R. Stat. Soc. Ser. B
**1951**, 13, 1–38. [Google Scholar] [CrossRef] - Khuri, A.I.; Mukhopadhyay, S. Response surface methodology. Wiley Interdiscip. Rev. Comput. Stat.
**2010**, 2, 128–149. [Google Scholar] [CrossRef] - Sarabia, L.A.; Ortiz, M.C. Response Surface Methodology. In Comprehensive Chemometrics; Elsevier: Amsterdam, The Netherlands, 2009; pp. 345–390. [Google Scholar]
- Wu, J.C.F.; Hamada, M.S. Experiments: Planning, Analysis, and Optimization, 3rd ed.; Wiley Series in Probability and Statistics; Wiley: New York, NY, USA, 2021; ISBN 9781119470106. [Google Scholar]
- de Oliveira, L.G.; de Paiva, A.P.; Balestrassi, P.P.; Ferreira, J.R.; da Costa, S.C.; da Silva Campos, P.H. Response surface methodology for advanced manufacturing technology optimization: Theoretical fundamentals, practical guidelines, and survey literature review. Int. J. Adv. Manuf. Technol.
**2019**, 104, 1785–1837. [Google Scholar] [CrossRef] - Karimifard, S.; Alavi Moghaddam, M.R. Application of response surface methodology in physicochemical removal of dyes from wastewater: A critical review. Sci. Total Environ.
**2018**, 640–641, 772–797. [Google Scholar] [CrossRef] - Myers, R.H.; Montgomery, D.C.; Anderson-Cook, C.M. Response Surface Methodology: Process and Product Optimization Using Designed Experiments, 3rd ed.; John Wiley & Sons: New York, NY, USA, 2016; ISBN 978-1-118-91601-8. [Google Scholar]
- Jankovic, A.; Chaudhary, G.; Goia, F. Designing the Design of Experiments (DOE)—An investigation on the influence of different factorial designs on the characterization of complex systems. Energy Build.
**2021**, 250, 111298. [Google Scholar] [CrossRef] - Carmignani, A.; Battaglini, M.; Sinibaldi, E.; Marino, A.; Vighetto, V.; Cauda, V.; Ciofani, G. In Vitro and Ex Vivo investigation of the effects of polydopamine nanoparticle size on their antioxidant and photothermal properties: Implications for biomedical applications. ACS Appl. Nano Mater.
**2022**, 5, 1702–1713. [Google Scholar] [CrossRef] - Guerrini, L.; Alvarez-Puebla, R.A.; Pazos-Perez, N. Surface modifications of nanoparticles for stability in biological fluids. Materials
**2018**, 11, 1154. [Google Scholar] [CrossRef] - Box, G.E.P.; Draper, N.R. Response Surfaces, Mixtures, and Ridge Analyses, 2nd ed.; Wiley: Hoboken, NJ, USA, 2007; ISBN 978-0-470-05357-7. [Google Scholar]
- Lenth, R.V. Response-Surface Methods in R., Using Rsm. J. Stat. Softw.
**2009**, 32, 1–17. [Google Scholar] [CrossRef] - Zhang, Z.; Xiaofeng, B. Comparison about the Three Central Composite Designs with Simulation. In Proceedings of the 2009 International Conference on Advanced Computer Control, Singapore, 22–24 January 2009; pp. 163–167. [Google Scholar]
- Naegle, K.; Gough, N.R.; Yaffe, M.B. Criteria for biological reproducibility: What does “n” mean? Sci. Signal.
**2015**, 8, fs7. [Google Scholar] [CrossRef] [PubMed] - Breig, S.J.M.; Luti, K.J.K. Response surface methodology: A review on its applications and challenges in microbial cultures. Mater. Today Proc.
**2021**, 42, 2277–2284. [Google Scholar] [CrossRef] - Gilman, J.; Walls, L.; Bandiera, L.; Menolascina, F. Statistical design of experiments for synthetic biology. ACS Synth. Biol.
**2021**, 10, 1–18. [Google Scholar] [CrossRef] [PubMed] - Klichinsky, M.; Ruella, M.; Shestova, O.; Lu, X.M.; Best, A.; Zeeman, M.; Schmierer, M.; Gabrusiewicz, K.; Anderson, N.R.; Petty, N.E.; et al. Human chimeric antigen receptor macrophages for cancer immunotherapy. Nat. Biotechnol.
**2020**, 38, 947–953. [Google Scholar] [CrossRef] - Xiong, R.; Vandenbroucke, R.E.; Broos, K.; Brans, T.; Van Wonterghem, E.; Libert, C.; Demeester, J.; De Smedt, S.C.; Braeckmans, K. Sizing Nanomaterials in bio-fluids by CFRAP enables protein aggregation measurements and diagnosis of bio-barrier permeability. Nat. Commun.
**2016**, 7, 12982. [Google Scholar] [CrossRef] [PubMed] - Seber, G.A.F.; Lee, A.J. Linear Regression Analysis; Wiley Series in Probability and Statistics; Wiley: New York, NY, USA, 2003; ISBN 9780471415404. [Google Scholar]
- Alcantara, I.M.; Naranjo, J.; Lang, Y. Model selection using press statistic. Comput. Stat.
**2022**, 37, 1–14. [Google Scholar] [CrossRef] - Bar-Gera, H. The target parameter of adjusted r-squared in fixed-design experiments. Am. Stat.
**2017**, 71, 112–119. [Google Scholar] [CrossRef] - Chicco, D.; Warrens, M.J.; Jurman, G. The coefficient of determination r-squared is more informative than SMAPE, MAE, MAPE, MSE and RMSE in regression analysis evaluation. PeerJ Comput. Sci.
**2021**, 7, e623. [Google Scholar] [CrossRef] - Eriksson, L.; Jaworska, J.; Worth, A.P.; Cronin, M.T.D.; McDowell, R.M.; Gramatica, P. Methods for reliability and uncertainty assessment and for applicability evaluations of classification- and regression-based QSARs. Environ. Health Perspect.
**2003**, 111, 1361–1375. [Google Scholar] [CrossRef] [PubMed] - Donev, A.N. Design of experiments in the presence of errors in factor levels. J. Stat. Plan. Inference
**2004**, 126, 569–585. [Google Scholar] [CrossRef] - Baş, D.; Boyacı, İ.H. Modeling and optimization I: Usability of response surface methodology. J. Food Eng.
**2007**, 78, 836–845. [Google Scholar] [CrossRef] - Kościelniak, P.; Parczewski, A. Non-linear transformation of factors in the design of experiments. Anal. Chim. Acta
**1983**, 153, 103–109. [Google Scholar] [CrossRef] - Zolgharnein, J.; Shahmoradi, A.; Ghasemi, J.B. Comparative Study of box-behnken, central composite, and doehlert matrix for multivariate optimization of Pb (II) adsorption onto robinia tree leaves. J. Chemom.
**2013**, 27, 12–20. [Google Scholar] [CrossRef] - Duckert, B.; Fauvart, M.; Goos, P.; Stakenborg, T.; Lagae, L.; Braeken, D. High-definition electroporation: Precise and efficient transfection on a microelectrode array. J. Control. Release
**2022**, 352, 61–73. [Google Scholar] [CrossRef] - Box, G.E.P.; Hunter, J.S.; Hunter, W.G. Statistics for Experimenters: Design, Innovation, and Discovery, 2nd ed.; Wiley: Hoboken, NJ, USA, 2005; ISBN 978-0-471-71813-0. [Google Scholar]
- Wolf, T.; Jin, W.; Zoppi, G.; Vogel, I.A.; Akhmedov, M.; Bleck, C.K.E.; Beltraminelli, T.; Rieckmann, J.C.; Ramirez, N.J.; Benevento, M.; et al. Dynamics in protein translation sustaining T cell preparedness. Nat. Immunol.
**2020**, 21, 927–937. [Google Scholar] [CrossRef] - Howden, A.J.M.; Hukelmann, J.L.; Brenes, A.; Spinelli, L.; Sinclair, L.V.; Lamond, A.I.; Cantrell, D.A. Quantitative analysis of T cell proteomes and environmental sensors during T cell differentiation. Nat. Immunol.
**2019**, 20, 1542–1554. [Google Scholar] [CrossRef] - Divecha, J.; Tarapara, B. Small, balanced, efficient, optimal, and near rotatable response surface designs for factorial experiments asymmetrical in some quantitative, qualitative factors. Qual. Eng.
**2017**, 29, 196–210. [Google Scholar] [CrossRef] - Wu, C.F.J.; Ding, Y. Construction of response surface designs for qualitative and quantitative factors. J. Stat. Plan. Inference
**1998**, 71, 331–348. [Google Scholar] [CrossRef] - Dulbecco, R.; Vogt, M. Plaque formation and isolation of pure lines with poliomyelitis viruses. J. Exp. Med.
**1954**, 99, 167–182. [Google Scholar] [CrossRef] [PubMed] - Rodgers, J.L.; Nicewander, W.A.; Toothaker, L. Linearly independent, orthogonal, and uncorrelated variables. Am. Stat.
**1984**, 38, 133–134. [Google Scholar] [CrossRef] - Rao, A.C.; Toutenburg, H.; Shalabh; Heumann, C. Linear Models and Generalizations; Springer Series in Statistics; Springer: Berlin/Heidelberg, Germany, 2008; ISBN 978-3-540-74226-5. [Google Scholar]
- Willmott, C.; Matsuura, K. Advantages of the mean absolute error (MAE) over the root mean square error (RMSE) in assessing average model performance. Clim. Res.
**2005**, 30, 79–82. [Google Scholar] [CrossRef] - Yap, B.W.; Sim, C.H. Comparisons of various types of normality tests. J. Stat. Comput. Simul.
**2011**, 81, 2141–2155. [Google Scholar] [CrossRef] - Uyanto, S.S. Monte carlo power comparison of seven most commonly used heteroscedasticity tests. Commun. Stat.-Simul. Comput.
**2022**, 51, 2065–2082. [Google Scholar] [CrossRef] - Tillman, J.A. The power of the durbin-watson test. Econometrica
**1975**, 43, 959. [Google Scholar] [CrossRef] - Buyske, S.; Trout, R. Advanced Design of Experiments. Available online: https://statweb.rutgers.edu/buyske/591/lect_all.pdf (accessed on 20 September 2022).
- Axler, S. Linear Algebra Done Right. In Undergraduate Texts in Mathematics, 3rd ed; Springer International Publishing: Cham, Switzerland, 2015; ISBN 978-3-319-11079-0. [Google Scholar]
- Li, J.J. Spectral Theorem and Applications; University of Chicago: Chicago, IL, USA, 2016. [Google Scholar]
- Lankham, I.; Nachtergaele, B.; Schilling, A. The Spectral Theorem for Normal Linear Maps. University of California: Los Angeles, CA, USA, 2007. [Google Scholar]
- Halmos, P.R. What does the spectral theorem say? Am. Math. Mon.
**1963**, 70, 241. [Google Scholar] [CrossRef] - NIST/SEMATECH. NIST/SEMATECH E-Handbook of Statistical Methods; NIST/SEMATECH: Gaithersburg, MD, USA, 2003. [Google Scholar]

**Figure 1.**Characterization of BSA-coated and uncoated polydopamine nanoparticles (PDNPs). (

**A**) Evolution of the PDNP hydrodynamic diameter in the function of reaction time as measured by DLS. All samples were sonicated by tip sonication to remove potential agglomerates. (

**B**) Normalized UV/Vis spectrum of DA (blue), BSA (red), 500 nm PDNPs (dashed blue) and 500 nm BSA-coated PDNPs (dashed red). (

**C**) Size distributions representing the stability of uncoated 500 nm PDNPs in water (blue) and Opti-MEM (dashed blue), showing a clear size increase after incubation in Opti-MEM. BSA-coated PDNPs maintain the same size distribution in both waters (red) and Opti-MEM (dashed red). Measured by DLS for an incubation time of 15 min.

**Figure 2.**Size and morphology characterization of BSA-coated polydopamine nanoparticles (PDNPs) of various sizes. (

**A**–

**C**) Representative intensity size distributions as measured by DLS of BSA-coated PDNPs of 300 nm (

**A**), 500 nm (

**B**) and 700 nm (

**C**) in water (blue) and 15 min incubated in Opti-MEM (red). (

**D**–

**F**) Representative SEM images of BSA-coated PDNPs of 300 nm (

**D**), 500 nm (

**E**) and 700 nm (

**F**). Scale bar = 500 nm. (

**G**–

**I**) Representative TEM images of BSA-coated PDNPs of 300 nm (

**G**), 500 nm (

**H**) and 700 nm (

**I**). Scale bar = 500 nm.

**Figure 3.**VNB formation by PDNPs. (

**A**) Visualization of vapor nanobubbles (VNBs) (red arrows) by dark field microscopy before (pre) and upon (pulse) local laser irradiation with a fluence of 1 J/cm

^{2}in a 500 nm PDNP dispersion (dashed yellow circle). (

**B**–

**F**) VNB formation as a function of laser pulse fluence for 142 nm (

**B**), 300 nm (

**C**), 500 nm (

**D**), 700 nm (

**E**) and 858 nm (

**F**) PDNPs. By fitting a Boltzmann curve (solid black curve), the VNB threshold (dashed red line) could be determined for 500 nm, 700 nm and 858 nm PDNPs. Scale bar = 50 µm.

**Figure 4.**Representative response surfaces where the PDNP size was kept constant, and the other factors were varied. (

**A**) CCC design-based model output visualized, showing a clear optimum. (

**B**) BBD-based model output visualized, with no clear optimum being present in the investigated design space. (

**C**) Revised BBD-based model output visualized, now showing a clear optimum.

**Figure 5.**Comparison of the delivery yield experimentally obtained at different factor level combinations. Yield in the function of different PDNP sizes, PDNP concentrations and laser fluences, showing the presence of unfavorable factor level combinations. Near-optimal parameter value combinations led to delivery yields close to the mean (±standard deviation) of the predicted optimal delivery yield by the CCC design- (dotted line; 39.168 ± 2.965%) and revised BBD-based (alternating stripe and dot line; 44.577 ± 1.971%) models. In the case of the BBD, these values were near the mean of its predicted stationary point by BBD-based models (striped line; 40.885 ± 2.763%). Error bars = standard deviations.

**Figure 6.**Scheme representing the complete RSM procedure used in this study and clarifying the model validation strategy.

**Table 1.**Summary of the average (±standard deviation) hydrodynamic diameter (size), polydispersity index (PDI), and zeta potential (Z.P.) of the BSA coated PDNPs with sizes corresponding to factor levels −1, 0 and 1 in both HyClone Pure water and Opti-MEM. Z.P. was only measured for BSA-coated PDNPs in HyClone Pure water.

300 nm Factor Level: −1 | 500 nm Factor Level: 0 | 700 nm Factor Level: +1 | ||||
---|---|---|---|---|---|---|

HyClone | Opti-MEM | HyClone | Opti-MEM | HyClone | Opti-MEM | |

Size (nm) | 290.8 ± 3.5 | 309.0 ± 6.8 | 516.1 ± 9.3 | 492.3 ± 5.8 | 728.4 ± 8.7 | 754.6 ± 23.6 |

PDI | 0.084 ± 0.020 | 0.078 ± 0.015 | 0.051 ± 0.014 | 0.050 ± 0.019 | 0.118 ± 0.048 | 0.112 ± 0.043 |

Z.P. (mV) | −33.0 ± 0.1 | −37.9 ± 0.9 | −40.6 ± 0.5 |

**Table 2.**Summary statistics for the optimization of photoporation yield using PDNPs as photothermal NPs and FD500 as cargo molecules, with both a CCC design and BBD. Triplicates were performed to assess model output reproducibility. The mean and standard deviation of the triplicates are presented. Values indicated in red and with * are significant at p < 0.05 for a one-sided unpaired Welch’s t-test between the CCC design- and BBD-based model summary statistics.

CCC | BBD | |
---|---|---|

Statistic | Mean ± Standard Deviation | Mean ± Standard Deviation |

Multiple R^{2} | 0.871 ± 0.038 * | 0.951 ± 0.009 * |

Adjusted R^{2} | 0.726 ± 0.081 * | 0.877 ± 0.022 * |

PRESS | 1921 ± 821 | 907 ± 79 |

Predicted R^{2} | 0.145 ± 0.275 | 0.416 ± 0.046 |

RMSE | 3.957 ± 0.824 * | 2.180 ± 0.191 * |

MAE | 3.340 ± 0.505 * | 1.850 ± 0.185 * |

**Table 3.**Summary of the stationary points (S.P.) of all replicates for both the CCC design and BBD in original units, delivery yield at the S.P., coefficients of variation (C.V.) and all corresponding eigenvalues, providing more information on the nature of the S.P.

CCC | BBD | |||||||
---|---|---|---|---|---|---|---|---|

Replicate 1 | Replicate 2 | Replicate 3 | C.V. | Replicate 1 | Replicate 2 | Replicate 3 | C.V. | |

Size (nm) | 534.956 | 550.250 | 536.688 | 1.549 | 524.159 | 540.399 | 567.548 | 4.029 |

Concentration (×10^{8} NPs/mL) | 2.594 | 2.455 | 2.504 | 2.800 | 2.823 | 2.404 | 1.279 | 36.817 |

Fluence (J/cm^{2}) | 0.950 | 0.841 | 0.939 | 6.594 | 1.076 | 0.745 | 0.793 | 20.528 |

Delivery yield at S.P. (%) with 95% confidence intervals | 41.899 ± 7.044 | 39.591 ± 7.684 | 36.015 ± 5.214 | 7.569 | 37.781 ± 3.981 | 43.076 ± 4.515 | 41.798 ± 8.796 | 6.758 |

Eigenvalues | −2.393 | −2.053 | −2.718 | N.A. | −1.491 | 0.693 | 0.915 | N.A. |

−3.822 | −4.777 | −4.169 | −3.494 | −0.956 | −0.851 | |||

−10.576 | −10.385 | −10.025 | −16.884 | −18.166 | −17.735 |

**Table 4.**Summary of the unpaired Welch’s t-test between the averaged predicted values and averaged observed values for the BBD-based models, with standard deviations (SD) also reported. Values indicated in red and with * are significant at p < 0.05.

Variables | Observed Values | Predicted Values | t-Test | ||||||
---|---|---|---|---|---|---|---|---|---|

Run | Size | Concentration | Fluence | Mean | SD | Mean | SD | t-Value | p-Value |

1 | 0.25 | 0.5 | 0.5 | 29.437 | 1.902 | 38.215 | 1.973 | −5.548 | 0.005 * |

2 | −0.5 | −0.5 | 0.5 | 24.316 | 7.639 | 33.033 | 4.837 | −1.670 | 0.183 |

3 | −0.5 | 0.5 | −0.5 | 24.166 | 5.488 | 33.613 | 3.379 | −2.539 | 0.077 |

4 | 0.25 | −0.5 | −0.5 | 39.794 | 4.459 | 39.713 | 4.606 | 0.022 | 0.984 |

5 | 0.25 | 0.5 | −0.5 | 34.062 | 1.541 | 39.812 | 2.952 | −2.991 | 0.058 |

6 | 0.25 | −0.5 | 0.5 | 32.018 | 4.620 | 39.429 | 3.919 | −2.119 | 0.103 |

7 | −0.5 | 0.5 | 0.5 | 28.528 | 5.513 | 33.866 | 2.278 | −1.550 | 0.230 |

8 | −0.5 | −0.5 | −0.5 | 20.523 | 3.739 | 31.467 | 4.622 | −3.188 | 0.035 * |

9 | −1 | 1 | 1 | 19.168 | 7.338 | 20.149 | 2.872 | −0.216 | 0.845 |

10 | 1 | 0 | 0 | 26.262 | 4.842 | 26.630 | 1.628 | −0.125 | 0.910 |

11 | −1 | 1 | −1 | 19.090 | 7.274 | 18.492 | 7.245 | 0.101 | 0.925 |

12 | −1 | −1 | 1 | 13.305 | 5.476 | 17.069 | 8.981 | −0.620 | 0.576 |

13 | 0 | 0 | 1 | 23.111 | 8.938 | 38.893 | 2.800 | −2.919 | 0.081 |

14 | 0 | 0 | −1 | 38.841 | 2.164 | 39.541 | 4.478 | −0.244 | 0.824 |

15 | 1 | 1 | 1 | 20.239 | 3.217 | 17.903 | 3.677 | 0.828 | 0.455 |

16 | 1 | 1 | −1 | 28.705 | 1.443 | 26.108 | 5.100 | 0.849 | 0.475 |

17 | −1 | −1 | −1 | 15.425 | 2.595 | 10.159 | 6.440 | 1.314 | 0.292 |

18 | −1 | 0 | 0 | 22.932 | 9.079 | 18.488 | 2.998 | 0.805 | 0.492 |

19 | 0 | −1 | 0 | 34.199 | 10.266 | 38.359 | 6.726 | −0.587 | 0.593 |

20 | 0 | 1 | 0 | 37.972 | 5.081 | 38.609 | 2.644 | −0.193 | 0.860 |

21 | 1 | −1 | −1 | 34.704 | 3.256 | 28.689 | 7.757 | 1.238 | 0.313 |

22 | 1 | −1 | 1 | 28.532 | 3.118 | 25.737 | 6.291 | 0.690 | 0.541 |

**Table 5.**Overview of the parameter values included in the design and their translation to coded values (bold) as they are used in the RSM model.

Parameter Values | |||||
---|---|---|---|---|---|

Size (nm) | 142 | 300 | 500 | 700 | 858 |

Concentration (NPs/mL) | 0.789 × 10^{8} | 1.5 × 10^{8} | 2.5 × 10^{8} | 3.5 × 10^{8} | 4.289 × 10^{8} |

Fluence (J/cm^{2}) | 0.463 | 0.7 | 1 | 1.3 | 1.537 |

Coded factor levels | −1.789 | −1 | 0 | +1 | +1.789 |

**Table 6.**Overview of the parameter values included in the confirmatory runs and their conversion to coded values (bold) as they are used in the RSM model. * Coded factor level for a size of 550 nm is 0.25, while the other parameter values in that column correspond to coded factor levels of 0.5.

Parameter Values | ||
---|---|---|

Size (nm) | 400 | 550 |

Concentration (NPs/mL) | 2.0 × 10^{8} | 3.0 × 10^{8} |

Fluence (J/cm^{2}) | 0.85 | 1.15 |

Coded factor levels | −0.5 | 0.25 or 0.5 * |

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |

© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Goemaere, I.; Punj, D.; Harizaj, A.; Woolston, J.; Thys, S.; Sterck, K.; De Smedt, S.C.; De Vos, W.H.; Braeckmans, K.
Response Surface Methodology to Efficiently Optimize Intracellular Delivery by Photoporation. *Int. J. Mol. Sci.* **2023**, *24*, 3147.
https://doi.org/10.3390/ijms24043147

**AMA Style**

Goemaere I, Punj D, Harizaj A, Woolston J, Thys S, Sterck K, De Smedt SC, De Vos WH, Braeckmans K.
Response Surface Methodology to Efficiently Optimize Intracellular Delivery by Photoporation. *International Journal of Molecular Sciences*. 2023; 24(4):3147.
https://doi.org/10.3390/ijms24043147

**Chicago/Turabian Style**

Goemaere, Ilia, Deep Punj, Aranit Harizaj, Jessica Woolston, Sofie Thys, Karen Sterck, Stefaan C. De Smedt, Winnok H. De Vos, and Kevin Braeckmans.
2023. "Response Surface Methodology to Efficiently Optimize Intracellular Delivery by Photoporation" *International Journal of Molecular Sciences* 24, no. 4: 3147.
https://doi.org/10.3390/ijms24043147