Gene Transfection of Primary Mouse Hepatocytes in 384-Well Plates

Abstract

We report the development of an in vitro transfection assay to test the efficiency of non-viral vector DNA nanoparticle transfection of primary hepatocytes. The protocol describes the isolation of viable hepatocytes from a mouse by collagenous perfusion. Primary mouse hepatocytes are plated in 384-well plates and cultured for 24 hours prior to transfection with polyethylenimine (PEI) or peptide DNA nanoparticles. Luciferase expression is measured after 24 hours following the addition of ONE-Glo substrate. The gene transfer assay for primary hepatocytes was optimized for cell plating number, DNA dose, and PEI to DNA ratio. The assay was applied to compare the expression mediated by mRNA relative to two plasmids possessing different promoters. The reported assay provides reliable in vitro expression results that allow direct comparison of the efficiency of different non-viral gene delivery vectors.

Introduction

The successful development of non-viral vectors to achieve liver transfection following intravenous dosing of DNA nanoparticles depends upon many factors¹. The nanoparticle must be sufficiently stable and biocompatible with blood^2–3. The particles must be of sufficiently small size and possess a charge masking system that allows nanoparticles to avoid uptake by Kupffer cells^4–5. DNA nanoparticles must include an endosomal escape mechanism to allow particles to gain access to the cytosol and to avoid lysosomal destruction⁶. DNA nanoparticles are often endowed with targeting ligands, such as glycoconjugates, which facilitate cell entry via receptor mediated endocytosis via the asialoglycoprotein receptor (ASGP-R)^7–9. Most DNA particle properties described above can be optimized using in vivo biodistribution experiments combined with HepG2 cell mediated in vitro gene transfer studies^{3, 5, 10–11}. However, HepG2 cells possess many fewer ASGP-Rs compared to primary hepatocytes¹². Likewise, HepG2 cells are rapidly dividing which limits their utility for studying nuclear uptake of DNA nanoparticles in quiescent hepatocytes. Consequently, there is a need for a reliable 384-well gene transfer assay for primary hepatocytes to provide a means to test new non-viral vectors for in vitro gene transfer properties.

To develop a high throughput gene transfer assay, we have previously reported the miniaturization of transfection of HepG2 and primary hepatocytes into 384 and 1536-well assays¹⁰. The studies took advantage of polyethylenimine (PEI) as a potent in vitro gene transfer agent¹³. PEI binds to and collapses DNA to form cationic nanoparticles of approximately 80 nm in diameter that spontaneous enter cells by pinocytosis¹⁴. PEI is a potent gene transfer agent because it buffers endosomes via the “proton sponge effect”, causing rapid swelling and endosomal burst to release DNA nanoparticles into the cytosol¹⁵. The rapid mitosis of HepG2 cells grown in culture allows DNA nanoparticles to gain greater access to the nucleus compared to primary hepatocytes, resulting in higher levels of expression¹⁰.

In the present study we have improved upon the primary hepatocyte transfection by optimizing hepatocyte viability following seeding in 384-well plates. A robust primary hepatocyte transfection assay was developed based on luciferase readout and compared to parallel transfections in HepG2 cells. Virtually any plasmid DNA can be substituted and packaged and delivered to primary hepatocytes under the protocol established. Comparison of several non-viral vectors under different gene transfer conditions suggest that the optimized primary hepatocyte transfection assay may also rely on hepatocyte mitosis.

Materials and Methods

Isolation of Primary Mouse Hepatocytes

ICR male mice (20 ± 5g) aged 5–12 weeks of age were purchased from Envigo Labs. Primary mouse hepatocytes were isolated with two-step collagenase perfusion as described with only slight modification from the published report¹⁶. Mice were anesthetized by i.p. injection of ketamine (100mg/kg) and xylazine (10mg/kg). A longitudinal abdominal incision was applied to expose heart. The intestines and other organs were displaced exposing the portal vein, inferior vena cava (IVC) and liver. The portal vein was cannulated using 25-gauge needle and immediately perfused with perfusion buffer-I (oxygenated Hank’s Balanced Salt solution (HBSS) without Ca²⁺ and Mg²⁺, Gibco with 0.5 mM EDTA), at a flow rate of 5 ml/min at 37°C. Successful cannulation resulted in swelling of liver and its blanching. Once this is observed, the IVC was cut to allow the excess fluid to drain. After 7 min the perfusion is switched to buffer-II (oxygenated Dulbecco’s Modified Eagle’s Medium (DMEM), Gibco) supplemented with 200 U/ml collagenase Type I (Worthington Biochemical) and perfused for another 8 min at 5 ml/min. The liver lobes were massaged with a cotton swab until the end of perfusion at which time the liver appears white and loses shape when pricked with forceps. The liver was removed and placed in Petri dish containing perfusion buffer-II and incubated at 4°C. Using forceps, the outer layer of Glisson’s capsule was removed to liberate the cells. The cells were re-suspended in 35 ml of DMEM and filtered through 70 μm mesh filter (Corning). The filtrate was centrifuged at 50 × g for 1 min at 4°C. The supernatant was discarded and the pellet was washed twice with DMEM to remove dead cells and debris. The cell pellet was re-suspended in 10 ml of DMEM supplemented with 2 wt/vol% BSA (Sigma-Aldrich) and centrifuged at 50 × g for 1 min at 4°C. The cell viability was measured using 0.4% trypan blue exclusion assay¹⁷. Since trypan blue overestimates viability, only cells that looked small, clear bright and which excluded trypan blue dye were counted as viable¹⁸. On an average a single mouse yielded 15–20 million cells with viability > 90%.

Plating Primary Hepatocytes in 384 Well Plates:

Primary mouse hepatocytes were diluted with plating/thawing media (Gibco) that was combined with Williams E Media (without phenol red, Gibco) and supplemented with 5 v/v% fetal bovine serum (FBS) as described by the vendor to a concentration of 1 million cells per ml. The hepatocytes were plated on collagen-coated 384 well plates (Greiner Bio-One CELLSTAR plate, with cover, from VWR). Collagen coating was carried out by diluting Type I rat tail collagen (Corning) in 20 mM acetic acid to a final concentration of 2.8 μg/50 μL/well. The plates were incubated at 37°C overnight then washed with sterile PBS prior to use.

Freshly isolated hepatocytes in plating medium were maintained at 4 °C until plating in a 384 well plate. Cells ranging from 250–2000 cells/well were diluted with hepatocyte maintenance media (Gibco) combined with Williams E. Media (without phenol red, Gibco) supplemented with 10 v/v% FBS (Gibco) as described by the vendor, and plated in 384 well in a volume of 50 μl per well. The hepatocytes are covered and allowed to adhere for 24 hrs at 37 °C in a humidified 5% CO₂ incubator prior to transfection.

Transfection of Primary Hepatocytes:

Two plasmids were used study to gene transfer efficiency by measuring luciferase expression. gWiz-Luc (Promega, Madison, WI, USA) is a 6.7 kbp plasmid encoding luciferase and possessing a CMV promoter. A second plasmid, pGL3 control vector (Promega, Madison, WI, USA), possesses 5.3 kbp encoding luciferase and the SV40 promoter and enhancer. A third plasmid (80A mRNA) was used as template DNA to generate in vitro transcribed luciferase mRNA¹⁹.

PEI DNA nanoparticles were prepared as described previously at an N:P (Nitrogen to Phosphate) ratio ranging from 7–11 by mixing plasmid DNA (0.4 μg) with PEI (0.5–2.5 μg in 10 μl) in HBM buffer (5 mM HEPES and 2.7 M mannitol, pH 7.5), followed by incubation at RT for 30 min prior to transfection of cells¹⁰.

Direct transfection was performed by plating primary mouse hepatocytes in collagen coated 384-well plates and allowed these to adhere for 24 hrs. Cells were transfected with PEI-DNA and incubated for 24 hrs at 37°C in a humidified 5% CO₂ incubator. 384-well plates were cooled to RT and centrifuged at 1000 rpm for 1 min. ONE-Glo (Promega) was added (10 μl per well) and the bioluminescence was measured after 5 min on an Envision plate reader (Perkin-Elmer)¹⁰.

Indirect transfection was performed by pre-incubating primary mouse hepatocytes with PEI-DNA or PAcr-Mel-DNA for 1.5 hrs at 4°C with end over end rotation. The cell suspensions were then plated in collagen coated 384-well plates and incubated for 48 hrs at 37°C in a humidified 5% CO₂ incubator. 384-well plates were cooled to RT and centrifuged at 1000 rpm for 1 min. ONE-Glo (Promega) was added (10 μl per well) and the bioluminescence was measured after 5 min on an Envision plate reader (Perkin-Elmer)¹⁰.

Results and Discussion

The in vitro transfection of primary mouse hepatocytes in 384-well plates was examined in detail to develop a standardized protocol. There are many excellent examples of published protocols to isolate primary rat and mouse hepatocytes^{16, 20}. The collagenase perfusion method reported here is largely taken from a report by Alnylum who developed optimize parameters to study the in vitro delivery of targeted siRNA to primary mouse hepatocytes¹⁶. However, there are only sparse reports of isolated primary hepatocytes being used for gene delivery experiments^21–26. We have previously reported the transfection of primary mouse hepatocytes with PEI-DNA in a 384-well assay¹⁰ (Scheme 1). The refinement of that transfection protocol, along with improvements to increase its reliability, were necessary to report in the present study.

Scheme 1. Structures of PEI and PAcr-Mel Gene Transfer Agents.

PEI is a branched polyethylenimine of 25k Da average molecular weight. PAcr-Mel is composed of DNA binding polyacridine peptide linked by disulfide bond to a melittin membrane lytic peptide derivative.

The use of Gibco maintenance media as reported by Alnylum was essential to maintaining hepatocyte viability after plating¹⁶. Other media, such as Williams E. media, were examined but resulted greater loss of viability during the first 24 hour incubation following plating. The coating of 384-well plates with collagen proved to be of minor importance for achieving hepatocyte adhesion and subsequent transfection. In addition to these parameters, the use of collagenase type I from Worthington proved to be critical. Throughout the study, primary hepatocytes from different size mice ranging from 20–40 g were used without influencing the reliability of the assay read out.

The level of luciferase expression was examined as a function of the number of primary hepatocytes plated (Fig. 1). The transfection of primary hepatocytes was simultaneously analyzed as a function PEI:DNA (N:P) ratio (Fig. 1). Primary hepatocytes were reliably transfected with 400 ng of PEI-DNA when either 1000 or 2000 cells were plated per well. Even though maximal DNA compaction occurs at N:P of 2, maximal expression occurred at higher N:P ratios. The addition of free PEI to the cell culture media to achieve N:P 7 or 9 significantly increased the gene expression 100-fold by facilitating endosomal escape by the proton sponge effect¹⁵ (Fig. 1). Higher N:P ratios result in decreased expression due to PEI mediated cell toxicity²⁷.

Figure 1. PEI-DNA Transfection of Primary Hepatocytes.

Transfection of primary mouse hepatocytes with PEI-DNA (gWiz-Luc) at an N:P of 2 with 400 ng of gWiz-Luc with increasing free PEI from N:P 1–15. The luciferase relative light units (RLU) span a range of 1000 from a background of 10¹ to a maximum of 10⁴. Values presented are the mean ± S.D. of 6 samples per group. Statistical analysis was performed using two-way ANOVA (analysis of variance) within GraphPad Prism version 8.1.2 (GraphPad Software, CA, USA). * p<0.05; ** p<0.01; *** p<0.001.

To further optimize the PEI N:P ratio, the cell seeding density was varied from 250–2000 cells per well and PEI-DNA nanoparticles were prepared at N:P ratios of 7, 9 and 11. The results confirmed that maximal transfection occurred when delivering 400 ng of gWiz plasmid prepared at N:P 7 or 9 onto 1000 or 2000 primary hepatocytes per well seeded on collagen coated 384-well plates.

The gene transfer efficiency of mRNA was also compared with plasmids (gWiz-Luc and pGL3) that differ by either a CMV or SV40 promoter. Transfection of primary hepatocytes with an escalating dose of PEI-mRNA, PEI-gWiz-Luc and PEI-pGL3, at a PEI N:P ratio of 7 demonstrated that gWiz-Luc provides approximately a 100–500 fold increase in gene expression due to the CMV promotor (Fig. 3). By comparison, PEI-pGL3 and PEI-mRNA were both equivalent producing approximately 10-fold expression over background (Fig. 3). The dose response established that PEI-gWiz-Luc produced maximal expression at a dose of 200–400 ng (Fig. 3).

Figure 3.

Comparison of gWiz-Luc, pGL3 and mRNA delivered with PEI at N:P of 9 (2000 cells/well). Values presented are the mean ± S.D. of 6 samples per group. Statistical analysis was performed using two-way ANOVA (analysis of variance) within GraphPad Prism version 8.1.2 (GraphPad Software, CA, USA). * p<0.05; ** p<0.01; *** p<0.001 compared to DNA or mRNA dose of 0 ng.

Even though PEI is of tremendous value as a potent in vitro gene transfer agent, in most cases and depending on the route of delivery, it fails as an in vivo gene transfer vector¹³. PEI relies upon the proton sponge effect to mediated endosomal escape making it a challenge to generate sufficient free PEI concentrations in vivo to maximize expression. To overcome the limitations of PEI for in vivo gene delivery we have been developing low molecular peptides that form metabolically stable DNA nanoparticles that function in vivo. Advanced peptide designs such as PAcr-Mel possess a disulfide bond linking the polyacridine peptide to a melittin analogue with minimized positive charge^{6, 28} (Scheme 1). Upon pinocytosis of the DNA nanoparticle, the reduction of the disulfide bond releases melittin which forms trimers that effect endosomal lysis leading to the release of the DNA nanoparticles into the cytosol⁶.

Comparison of the gene transfer efficiency of PEI-DNA and PAcr-Mel-DNA in primary mouse hepatocytes and HepG2 cells established that while both transfected plated HepG2 cells with similar efficiency (Fig. 4A), only PEI-DNA was active when used to directly transfect plated primary hepatocytes (Fig. 4B). In an effort to increase PAcr-Mel-DNA transfection efficiency in primary hepatocytes an indirect protocol was developed by incubating nanoparticles and cells with end-over-end rotation at 4°C for 1.5 hrs prior to plating on collagen coated plates, followed by luciferase measurement at 48 hrs. This significantly increased the gene transfer from PAcr-Mel-DNA but also eliminated the gene transfer mediated by PEI-DNA (Fig. 4B).

Figure 4.

Comparison of PAcr-Mel versus PEI mediated gene transfer efficiency. Panel A illustrates equivalent gene transfer efficiency by direct transfection (added to dividing cells) with PEI-DNA and PAcr-Mel-DNA in HepG2 cells. Panel B illustrates a significant reduction in gene transfer by direct transfection of PAcr-Mel-DNA relative to PEI-DNA in primary hepatocytes. However, indirect transfection (premixing with cells prior to plating) resulted in a significant decrease in PEI-DNA versus PAcr-Mel-DNA mediated gene transfer. Values presented are the mean ± S.D. of 6 samples per group. Statistical analysis was performed using One sample t-test within GraphPad Prism version 8.1.2 (GraphPad Software, CA, USA). ** p<0.01.

Since the PAcr-Mel is a disulfide reducible gene transfer agent, we examined the glutathione in both HepG2 cells and primary hepatocytes^29–30. HepG2 cells possessed approximately 25% of glutathione determined for primary hepatocytes. How this relates to the differential gene transfer of PAcr-Mel in primary hepatocytes remains to be determined with further investigations.

In conclusion, the transfection of primary hepatocytes protocol described here is a refinement of our previous report which better defines important media and plating conditions that allow consistent hepatocyte transfections in 384-well plates. Since isolated primary hepatocytes are in a quiescent (G0) phase of cell division³¹, these cells are more difficult to transfect compared to hepatocyte cell lines such as HepG2. The reason for this difference is related to rapid cell division of cultured HepG2 cells which allows DNA nanoparticles to gain access to the nucleus. Plating primary hepatocytes in 384-well collagen coated plates followed by seeding for 24 hours in Gibco maintenance media at 37°C in a CO₂ incubator allowed quiescent hepatocytes to undergo cell division to facilitate nuclear uptake of PEI-DNA nanoparticles. The addition of the 24 hour plating step to allow hepatocytes to undergo division proved to be essential to increase assay reliability. However, while plating primary hepatocytes greatly improves PEI-DNA mediated gene transfer it fails to provide a transfection assay for PAcr-Mel-DNA which relies upon disulfide bound reduction. The indirect pre-mixing protocol described overcomes this by increasing the expression PAcr-Mel-DNA but does not support PEI-DNA expression. The reasons for these contradictory relationships between PEI-DNA and PAcr-Mel-DNA is presently unclear. With further refinement, the assay described will be useful to test the transfection efficiency of novel gene delivery systems in primary hepatocytes.

Figure 2.

Optimization of PEI N:P ratio from 7–11 with 400 ng of gWiz-Luc with different cell seeding number. Values presented are mean ± S.D. of 6 samples per group. Statistical analysis was performed using two-way ANOVA (analysis of variance) within GraphPad Prism version 8.1.2 (GraphPad Software, CA, USA). * p<0.05; ** p<0.01; *** p<0.001