pH Modification of Gel Mobility Shift Improves Polyplex Selection In Vivo

Abstract

Cationic polymers that bind with the plasmids to form polyplexes protect the DNA from enzymatic degradation and improve cellular and tissue uptake. Complete or near complete gel retardation of the polyplex is an important assay to determine the optimal polymer: plasmid ratio for in vitro and in vivo studies. Nevertheless, despite minimal to moderate gel retardation of histidine-lysine (HK) polyplexes formed with low peptide: plasmid DNA ratios (1:2 and 1:4; w:w), the polyplexes effectively targeted the tumor in vivo. To understand the lack of predictability of the initial gel mobility shift assays, we revisited the retardation and stability of polyplexes with these electrophoresis assays. Because the histidine component with a pKa of about 6.0 will have a greater positive charge and may bind plasmids with a higher affinity at lower pHs, we compared the retardation of the two HK polyplexes when the pH of the running buffer of the gel mobility shift assay was altered. Both HK polyplexes were retarded significantly more when the running buffer had a pH of 7.3 instead of the standard pH of 8.3. Indeed, the HK polyplexes formed at the 1:2 ratio showed complete retardation at pH 7.3. Consequently, while both HK polyplexes formed at these low ratios targeted the tumor, the polyplex formed with the 1:2 ratio had reduced tumor gene expression variability and lower lung and liver values. Thus, the selection of the optimal ratios for the linear HK and plasmid for transfection studies in vivo was improved with a running buffer pH of 7.3.

1. Introduction.

Naked plasmid DNA has been touted as a safe and effective method to deliver genes, as a vaccine, or to the lungs [1–3]. For lung delivery, naked plasmid DNA delivery likely has the most utility in pathological conditions with low mucous secretions [2,4]. Because of the rapid enzymatic degradation and uptake by liver endothelial cells, plasmids delivered systemically show little to no tissue expression, particularly when administered by the tail vein [5,6]. The naked plasmid method likely has lower immunogenicity and toxicity than DNA nanoparticles without liposomal or polymeric carriers [1]. Nevertheless, because a carrier is necessary for the stability and targeting of intravenously administered DNA, the lowest amount required for the delivery of DNA could arguably be ideal. We recently found that low concentrations of the polymeric carrier of plasmids given intravenously can target the tumor effectively [7].

Despite the efficient tumor delivery, gel retardation assay with these low polymers: plasmid DNA ratios displayed minimal retardation of the polyplex [7]. Indeed, most of the plasmid DNA of the polyplex migrated similarly with the naked plasmid DNA during electrophoresis. Notably, gel retention in the well and retardation of the polyplex has been a critical assay for determining the ratio of polymer to nucleic acid for in vivo studies [8,9]. Thus, the efficient in vivo delivery to tumors of this polyplex comprised of low amounts of the linear histidine-lysine peptide seemingly contradicts this conventional wisdom. One possible explanation is that the high pH of 8.3 for the running buffer, the standard for our lab, may cause low retention of these histidine-containing polyplexes in the wells. For instance, histidines with a pKa of about 6.0 on the linear polymer will incur a greater positive charge and may bind to the plasmid with higher affinity if the gel running buffer has a pH of 7.3 than 8.3 [10–12].

Consequently, we revisited the migration and stability of polyplexes with gel mobility shift assays at different pHs. The HK polyplexes comprised of peptide: DNA ratios (1:2 and 1:4) were retarded significantly more during electrophoresis when the running buffer had a pH of 7.3. This result was in marked contrast when the pH of the running buffer was 8.3, where minimal retardation of the polyplex occurred, particularly at the lowest peptide: plasmid ratio. For linear HK peptide carriers of plasmids, the gel mobility shift assay with a pH running buffer of 7.3 was more predictive of targeting the tumor in vivo.

2. Materials and Methods.

2.1. Cell line

The cancer cell line, MDA-MB-231, was cultured and maintained in Dulbecco’s minimal essential medium (DMEM) containing 10% fetal calf serum (FCS) and 20 mM glutamine.

2.2. Peptides

The linear H2K peptide was synthesized on a peptide synthesizer (Genscript, Piscataway, NJ). To ensure 90% or greater purity, peptides were analyzed by high-performance liquid chromatography and further analyzed by mass spectroscopy. The sequence of H2K is as follows: H2K, KHKHHKHHKHHKHHKHHKHK (2.697 kDa).

2.3. Animals

Female athymic mice (4–8 weeks old) were purchased from Envigo (Indianapolis, IN). The Institutional Animal Care and Use Committee of the University of Maryland, Baltimore, approved the animal experiments.

2.4. Plasmids

pCpG-Luc (luciferase-expressing) plasmids were purified with an EndoFree Maxi kit (Qiagen, Valencia, CA) as described previously [13]. Gel electrophoresis was done on the luciferase plasmids to ensure that at least 60% of the plasmid was condensed. Endotoxin levels were measured per the manufacturer’s instructions to ensure low levels (Pierce Chromogenic Endotoxin Quant Kit-A39552S, Thermo Scientific-Waltham, MA). (Table 1S).

2.5. Tris-Acetate-EDTA (TAE), loading buffers, and pH modification

The pH of the TAE running buffer (1X, pH-8.3) (Quality Biological) and loading buffers (6X) were measured by the FiveEasy^™ meter; InLab Solids Pro-ISM pH electrode, Mettler Toledo, Columbus, OH, USA. Before pH measurements of the buffers, the pH standards (4.0, 7.0, and 10.0; Sigma-Aldrich, St. Louis, Mo) were done. In selected situations, the loading and running buffer pH was modified. The pH of the TAE (1X) running buffer was adjusted to 7.3 with 6 N HCl, while the pH of two loading buffers (NEB with or without SDS 6X, pH-8.5; New England Biolabs, Ipswich, MA) was adjusted with 1 N HCl. The different loading buffer dyes (6X) and their pH were the following: (1) Thermo (Thermo Fisher Scientific, Waltham, MA), pH-7.5; (2,3) NEB buffer without SDS, pH-8.5 and modified pH-7.4; (4,5) NEB buffer with SDS, pH-8.5 and modified pH-7.3; and (6) 40% sucrose, pH- 5.5.

2.6. Gel Mobility Shift Assays

Similar to in vivo studies (see below), the polyplexes were prepared at a peptide: plasmid DNA ratio of 1:2 (15 μg: 30 μg) or 1:4 (7.5 μg: 30 μg). Ten microliters of the polyplex and two microliters of various loading buffers were loaded in the well and electrophoresed on a 1 % gel for 25 to 60 minutes at 50 V. The pH of the TAE running buffer was 8.3 or 7.3.

2.7. Polyplex formation for in vivo use

To the plasmid (30 or 60 μg in 140 μl of 6.5 mM NaCl solution), the HK peptides (15 μg in 110 μl of 6.5 mM NaCl solution) were added and mixed by pipetting. The H3K peptide, which previously increased endosomal lysis, comprised 15% of the HK peptides [14].

2.8. Particle size, polydispersity index (PDI), and zeta potential (ZP)

As previously reported, the size (Z-average diameter (ZA)), PDI, and ZP of polyplexes were determined with the Zetasizer (Malvern, Westborough, MA) [12].

2.9. Tumor Xenografts and Injections

After MDA-MB-231 tumor cells (4 × 10⁶ cells) were injected bilaterally into the mammary fat pad and tumors were 150 mm³, the mice were injected in the tail vein with the plasmids or polyplexes (240 μl). To facilitate the injection of tail veins, the mice were warmed with a heating lamp for 10 min with the bottom of the container not exceeding 38°C. After 24 hours, the tumors and major organs were homogenized, and luciferase activity was measured. For tumor and normal tissues, luciferase activity was standardized by tissue weight. The number of tumor-bearing mice per group was as follows: naked plasmid (n=7 mice, 14 tumors), HK: plasmid 30 μg (n=6 mice, 12 tumors), HK plasmid 60 μg (n=6 mice, 11 tumors).

2.10. Statistical Analysis

The data were reported as Box Plots with the median, mean, and 10% and 90% error bars shown. Differences were compared using a one-way ANOVA of Ranks followed by a multiple comparisons Dunn post hoc test (Sigma Plot, version 15.0, Chicago, IL, USA). The gels with different running buffers and loading buffers were performed four times.

3. Results and Discussion.

3.1. The pH of the running buffer affects the retention of polyplexes.

Investigators have paid little attention to the effect of the loading and running buffer during gel electrophoresis on the stability of the nanoparticle. Complete or near complete retention of the polyplex within the well is a key criterion for determining the peptide-to-DNA ratio for in vitro and in vivo use. A prior publication by our group challenged this by finding histidine-containing polyplexes that were not retained in the well still had good tumor transfection in vivo [7]. As a result, we hypothesized that the lack of retention of these polyplexes may be partly because of the high pH of loading and running buffers on the pH-sensitive histidine-containing polyplexes. Moreover, the peptide’s histidine component with a pKa of about 6.0 would have a slightly greater positive charge and may bind to DNA more tightly with a buffer pH of 7.3 than 8.3 [10–12].

Consequently, we tested whether the loading and running buffers’ pH could affect the stability of polyplexes made up of pH-sensitive polymers such as the linear H2K peptide. In general, the loading buffer did not affect the stability of the polyplex. The only exception was the loading buffer with SDS (0.08%, 1X), which disrupted the polyplex completely (lanes 3, 5; Figures 1,2), while there was little to no effect of varying the pH of the loading buffers (lanes 2, 4; pH 8.3 vs pH 7.3; Figures 1, 2). There was greater retardation of the polyplex with the loading buffer from Thermo Scientific, but the effects were modest (lane 6; Figures 1,2).

Figure 1.

HK polyplexes were prepared at a peptide: plasmid ratio of 1:2 (15 μg: 30 μg). Plasmids or polyplexes (10 μl) mixed with various loading buffer dyes (LD) (2 μl) were loaded in the well and electrophoresed on a 1 % gel. pH of the TAE running buffer was 8.3 (A) and 7.3 (B). Samples were electrophoresed for 50 min at pH 8.3 and 60 min at pH 7.3. The plasmid control was in lane 1.

Figure 2.

HK polyplexes were prepared at a peptide: plasmid ratio of 1:4 (7.5 μg: 30 μg), loaded in the wells with different loading dyes, and electrophoresed on a 1 % gel. The pH of the TAE running buffer was 8.3 (A) and 7.3 (B).

In contrast, the pH of the running buffer significantly affected the polyplex’s stability. The excess volume of the running buffer compared to the loading buffer dictated the gel retardation and well retention of the polyplex. We determined that the polyplexes formed at different ratios (peptide: plasmid, 1:2; Figure 1; and peptide: plasmid ratio 1:4; Figure 2) were more stable with greater retardation when the running buffer had a pH of 7.3 instead of standard 8.3. At the higher peptide: plasmid ratio of 1:2, the polyplexes were completely retarded and retained in the wells at pH 7.3 (lanes 2, 4, 6, and 7; Figure 1B) than at pH 8.3 (Figure 1A). Moreover, the effect of the pH of the running buffer was particularly evident at the lower peptide: plasmid ratio (1:4; Figure 2). At this ratio, none of the polyplexes was retained in the well at the pH of 8.3 (lanes 2, 4, 6, and 7; Figure 2A). Accordingly, one would likely not further investigate the polyplexes formed at these low ratios in vivo. In contrast, the significant retention in the wells of the polyplexes with the pH of the running buffer of 7.3 would encourage further study in vivo (lanes 2, 4, 6, and 7; Figure 2B). As expected, the pH of the running buffer did not affect the migration of the naked plasmid (lanes 1; Figures 1, 2).

Many critical factors affect the stability of the polyplex in the circulation, such as the surface charge and pegylation. Nevertheless, the optimal ratio for the polyplex is also an important consideration for stability. Although complete or near retention in the well has been a criterion for selecting the optimal ratio, the rationale for selection may be more complex and nuanced. For the polyplex’s import into tumors in vivo, the optimal ratio is based on the nucleic acid (mRNA vs. plasmid), surface charge, the mechanism of entry (i.e., enhanced retention and permeation vs. neuropilin receptor-1 (NRP-1)), the degree of branching, and the pH dependency [15–18]. At least for pH-buffering polyplexes, considering the pH’s effect of the running buffer may lead to investigating peptide: plasmid ratios, which generally would be excluded from further study (Figure 1A), and improve prediction of the polyplex’s efficacy in vivo.

Although the modestly greater positive charge on the linear HK peptide at 7.3 stabilized the polyplex during gel electrophoresis, lowering the pH further may be disruptive because of charged-charge repulsion between the polymers. This charge repulsion may be more significant with the constrained branched than linear peptides. Stability and biophysical assays such as gel mobility shift assays, heparin displacement assays, and size measurements could investigate these possibilities further.

The question arises as to why the pH change of 8.3 to 7.3 would affect the histidine component with a pKa of about 6.0. This pH change would only alter the charge on the histidine component by about 10-fold (from 0.48% to 4.8%). However, the exact pKa of the histidine component of the peptide is unknown, and its pKa has been reported to range from 5.0 to 6.7 [12]. Although the pKa will not change the 10-fold difference, the protonated charge on histidines would differ dramatically from 1 to 10% with a pKa of 6.3. Moreover, the pKa of histidine can change in the presence of ions and nucleic acids [12,19,20]. Also, the protonated histidine charge will likely be stabilized, or the charge turnover will be much slower upon binding to a nucleic acid [19,20]. Finally, hydrogen bonding is strengthened between protonated histidines and nucleic acids as the pH is decreased [11]. All these factors may amplify the interactions between the HK peptide and the plasmid DNA at the lower pH.

3.2. Comparison of Polyplexes with “Naked” Plasmid DNA In Vivo

In addition, we compared the tumor and tissue expression of plasmids and polyplexes. In contrast to the plasmid-only group, both polyplexes had a similar size and PDI (Table 2S). Moreover, the tumor-targeting polyplexes gave markedly higher luciferase expression in the tumor than the plasmid-alone group (Figure 3). The HK peptide, with a -KHHK-sequence, targets the NRP1 transport system, which is highly expressed in many tumors; this transport system probably has the primary role in the polyplex’s uptake into the tumor [14,16,21]. Nevertheless, the stability of the polyplex, as indicated by the gel retardation, is likely a necessary but not essential requirement. With NRP-1 targeting and the significant gel retardation at pH 7.3, the polyplexes demonstrated markedly higher tumor expression than other tissues. The two polyplexes did not have statistical differences in the luciferase expression of their tumors, lungs, livers, or spleens.

Figure 3.

Box plot representation of luciferase-expressing tumor/tissues from mice injected intravenously with naked plasmid or two HK plasmid polyplex formulations. T, Tumor, Lu, lung, Li, liver, S, spleen. Median, solid line; Mean, dotted red line. **, P<0.01, Polyplex-tumor (30, 60 μg) > Plasmid-Tumor (60 μg), Anova of Ranks, Dunns Method for All Pairwise Multiple Comparisons Test. Lungs, liver, and spleen among the treatment groups were not significant.

Nevertheless, we prefer polyplexes containing 30 μg of plasmids (peptide: plasmid ratio 1:2) because the evidence suggests increased stability. These polyplexes’ greater stability and specificity were based on similar median and mean luciferase values in the tumors and lower lung and liver values (Figure 3, Table 1). Moreover, the greater specificity and stability suggested by this polyplex in vivo corresponded with the complete retardation of the polyplex at pH 7.3 (Figure 1). Still, the polyplex formed at the lower ratio (1:4) showed high tumor gene expression in vivo and demonstrated significantly greater gel retardation at pH 7.3 than 8.3 (Figures 2, 3). Additional ratios of peptide: plasmids for the retardation assay at pH 7.3 may yield more insight into polyplexes with improved targeting and specificity in vivo.

Table 1.

Median values of RLU per mg-protein from several tumors/tissues of mice injected intravenously with two different HK (15 μg) luciferase-expressing plasmid (30 or 60 μg) polyplexes or a “Naked” plasmid (60 μg). The mean and standard error are in parentheses.

Particles	Tumor	Lungs	Liver	Spleen
HK Polyplex (30 μg)	272.8 (253.2±57.5)	23.0 (41.6±28.7)	1.64 (1.9±0.7)	1.3 (1.4±0.5)
HK Polyplex (60 μg)	208.5 (311.7±72.8)	57.5 (85.5±32.3)	17.2 (44.5±28.1)	1.4 (1.6±0.4)
Plasmid (60 μg)	4.9 (47.7±26.7)	45.4 (151.4±60.3)	2.4 (48.4±30.5)	1.0 (2.1±0.8)

Surprisingly, the “naked” plasmid exhibited gene expression, particularly in the lungs, the first capillary organ encountered after injection. In contrast, other investigators have reported little to no gene expression in most organs administered plasmids intravenously [5,6]. Differences in tail vein injection methodology (i.e., total mouse vs. tail warming) or plasmid promoters are possible reasons for the dissimilar results. Notably, compared to polyplexes, the plasmid-only group had mean luciferase values that differed markedly from the median values of the tumor, lung, and liver tissues. Consequently, the luciferase values in these tumors/tissues were sporadic (Figure 3), consistent with the lack of protection from enzymatic degradation.

4. Conclusions.

The gel mobility shift assay is essential for determining the optimal peptide: plasmid ratio and the polyplexes’ stability. Because the histidine component of the carrier has a greater positive charge at lower pHs and may bind plasmids with a higher affinity, we compared the retardation of the polyplexes when the pH of the running or loading buffer was altered. Gel electrophoresis resulted in marked retardation of the HK polyplex at low peptide: plasmid ratios when the running buffer had a pH of 7.3 than the standard pH of 8.3. The loading buffer pH had little effect, indicating that the excess volume of the running buffer pH was the primary factor affecting the polyplex’s stability. Consequently, altering the pH of the running buffer for the gel mobility assay improved the selection of the optimal HK peptide: plasmid ratio for in vivo use. To our knowledge, this is the first report that has varied the loading or running buffers’ pH to investigate the polyplexes’ stability.

Highlights.

• Gel mobility shift assays with retardation of the polyplex are vital in determining the optimal polymer: plasmid ratio for in vivo studies.

• Lack of gel retardation of pH-sensitive HK polyplexes under standard conditions poorly correlates with in vivo transfection.

• HK polyplexes were retarded significantly more with a running buffer of pH 7.3 than with the standard pH 8.3.

• The HK polyplex, entirely retained in the well, showed higher specificity targeting the tumor in vivo.

• The selection of the optimal ratios for the linear HK polyplex for transfection studies in vivo was improved with a running buffer pH of 7.3.