The objective of this study was to compare different transgene expression and delivery strategies for expression of two transgenes in hMSCs and investigate transfection efficiency of each transgene. Specifically, we investigated four transgene expression and delivery strategies: (i) delivery of two DNA vectors, complexed separately, each expressing a single transgene; (ii) delivery of two DNA vectors, complexed together, each expressing a single transgene; (iii) delivery of a single DNA vector expressing two transgenes separated by an internal ribosome entry site (IRES) and (iv) delivery of a single DNA vector expressing two transgenes separated by a dual 2A peptide sequence (D2A); for their ability to express two reporter transgenes (enhanced green fluorescent protein [EGFP] and tandem dimer Tomato [tdTomato]) in four donors of hMSCs from two tissue sources using the commercially available transfection reagents Lipofectamine 3000 (lipid-mediated nonviral gene delivery) or Turbofect (polymer-mediated nonviral gene delivery) for complexation (Fig. 1). The effects of each transgene expression strategy on expression of both transgenes in hMSCs were assayed by fluorescence imaging of the expressed EGFP and tdTomato transgenes, normalized by total cell count (Hoechst 33342, nuclei stain), to obtain transfection efficiencies for all conditions. It is important to note that both mass of DNA delivered, and transgene copy number (i.e., molarity of transgene), were equal when directly comparing delivery of two DNA vectors to delivery of a single DNA vector that encodes for both transgenes to appropriately compare these conditions.

Fig. 1

Schematic of DNA Vectors, Conditions, and Experimental Design for this Study.  a  Representative schematics of DNA vectors used in this study along with approximate size of DNA vectors in kilobase pairs (kbp). Sequences for each DNA vector are available at https://www.addgene.org/Angela_Pannier/. White element: CMV promoter; red element: tdTomato; green element: EGFP; grey element: SV40 polyA signal; pink element: P2A-T2A; blue element: IRES. b  Conditions tested for expression of multiple transgenes in hMSCs. (i) Separate complex conditions consisted of forming complexes separately for each single transgene DNA vector (pEGFP or ptdTomato, denoted as [pE]+[pT]). (ii) Same complex conditions consisted of forming complexes with both single transgene DNA vectors together (pEGFP + ptdTomato, denoted as [pE + pT]). Bi-cistronic (iii) D2A and (iv) IRES DNA vectors were formed in individual complexes. All conditions had equal mass of DNA as well as copy number of transgenes when directly compared amongst each other. Mass and copy number were equalized by addition of a promoterless pEGFP plasmid where needed. c  hMSCs from four donors (D1, D2, D3, & D4) and two tissue sources (adipose and bone marrow; hAMSCs and hBMSCs, respectively) were transfected with the conditions shown in « b » 24 h after seeding (4,500 and 6,000 cells/well, respectively) and imaged for transfection efficiency of each transgene 24 h after transfection

Delivery of two DNA vectors each encoding a single transgene in hMSCs

We first established baseline transfection efficiencies for each transgene following transfection with a single DNA vector. It should be noted that an expressionless plasmid (i.e., promoter removed from expression cassette of pEGFP) was added to the single DNA vector conditions to equalize both mass of pDNA and moles of each transgene delivered for the conditions tested in Fig. 2 (see hMSC Transfection section in Materials and Methods for more detail). For single DNA vector conditions, transfection efficiency for pEGFP was approximately 31% with Lipofectamine 3000 and 44% with Turbofect, while transfection efficiency for ptdTomato was approximately 37% with Lipofectamine 3000 and 44% with Turbofect in D1 hAMSCs (Fig. 2a and b). For D2 hAMSCs, transfection efficiency for pEGFP was approximately 22% with Lipofectamine 3000 and 31% with Turbofect, while transfection efficiency for ptdTomato was approximately 23% with Lipofectamine 3000 and 32% with Turbofect (Fig. 2c and d). For D3 hBMSCs, transfection efficiency for pEGFP was approximately 37% with Lipofectamine 3000 and 56% with Turbofect, while transfection efficiency for ptdTomato was approximately 36% with Lipofectamine 3000 and 56% with Turbofect in D3 hBMSCs (Fig. 2e and f). Lastly, for D4 hBMSCs, transfection efficiency for pEGFP was approximately 28% with Lipofectamine 3000 and 41% with Turbofect, while transfection efficiency for ptdTomato was approximately 34% with Lipofectamine 3000 and 41% with Turbofect in D4 hBMSCs (Fig. 2g and h).

Fig. 2

Delivery of Multiple DNA Vectors in hMSCs for Expression of Two Reporter Transgenes.  hMSCs were transfected with single transgene vectors (pEGFP or ptdTomato) complexed with Lipofectamine 3000 (a, c, e, & g) or Turbofect (b, d, f, & h) and transfection efficiencies [i.e., number of EGFP (green bars), tdTomato (red bars), and co-expressing (both EGFP and tdTomato expression, yellow bars) cells relative to total cell counts] for each transgene (EGFP or tdTomato) were compared to two transgene delivery strategies; (i) two DNA vectors delivered as separate complexes ([pE]+[pT]); and (ii) two DNA vectors delivered in the same complex ([pE + pT]). Transfection efficiencies for hMSCs expressing both EGFP and tdTomato (co-expression) was calculated by dividing the number of cells that were both EGFP and tdTomato positive by the total cell count (Hoechst, nuclear stain). a Transfection efficiencies for all conditions in D1 hAMSCs that used Lipofectamine 3000 as the cationic transfection reagent. b Transfection efficiencies for all conditions in D1 hAMSCs that made use of Turbofect as the cationic transfection reagent. c Transfection efficiencies for all conditions in D2 hAMSCs that used Lipofectamine 3000 as the cationic transfection reagent. d Transfection efficiencies for all conditions in D2 hAMSCs that made use of Turbofect as the cationic transfection reagent. e Transfection efficiencies for all conditions in D3 hBMSCs that used Lipofectamine 3000 as the cationic transfection reagent. f Transfection efficiencies for all conditions in D3 hBMSCs that made use of Turbofect as the cationic transfection reagent. g Transfection efficiencies for all conditions in D4 hBMSCs that used Lipofectamine 3000 as the cationic transfection reagent. h Transfection efficiencies for all conditions in D4 hBMSCs that made use of Turbofect as the cationic transfection reagent. All conditions have equal moles of expression cassette and mass of DNA delivered. Data represented as mean ± SEM (n = 6). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001, ns, not significant (p > 0.05), as determined by a 2-way ANOVA with Tukey’s post hoc test

Once baseline transfection levels were measured for each single transgene, we examined if delivery of two separate DNA vectors, each encoding for a single transgene (i.e., EGFP or tdTomato; Fig. 1a), would result in expression of both transgenes in hMSCs by measuring transfection efficiencies (i.e., number of EGFP, tdTomato, and EGFP + tdTomato expressing cells divided by total cell counts) resulting from delivery of the vectors in two strategies: (i) two DNA vectors delivered as separate complexes (i.e., delivery of two DNA vectors, encoding for EGFP or tdTomato, in separate cationic polymer or lipid complexes; Fig. 1bi); and (ii) two DNA vectors delivered in the same complex (i.e., delivery of two DNA vectors, encoding for EGFP or tdTomato, mixed together prior to formation in the same cationic polymer or lipid complex; Fig. 1bii). The transfection efficiencies for the delivery of two separate DNA vectors were compared to the baseline transfection efficiencies for the single transgene DNA vectors measured above.

When two DNA vectors were delivered to D1 hAMSCs in separate complexes using Lipofectamine 3000 (denoted as “[pE]+[pT]”, Fig. 1bi), transfection efficiencies for EGFP and tdTomato were significantly reduced (adjusted p-value < 0.0001) to 18% and 23%, respectively, compared to the single transgene conditions (Fig. 2a). Similar results were seen when Turbofect was used as the cationic carrier to separately complex and deliver the two vectors, as transfection efficiencies for both EGFP and tdTomato were significantly reduced (adjusted p-value < 0.0001) to 31% and 34%, respectively, compared to the single transgene conditions (Fig. 2b). When two DNA vectors were delivered to D2 hAMSCs in separate complexes using Lipofectamine 3000, transfection efficiencies for EGFP and tdTomato were significantly reduced (adjusted p-value < 0.0001) to 12% and 14%, respectively, compared to the single transgene conditions (Fig. 2c). Similar results were seen when Turbofect was used as the cationic carrier to separately complex and deliver the two vectors, as transfection efficiencies for EGFP and tdTomato were significantly reduced (adjusted p-value < 0.0001) to 20% for both transgenes compared to the single transgene conditions (Fig. 2d). When two DNA vectors were delivered to D3 hBMSCs in separate complexes using Lipofectamine 3000, transfection efficiency for EGFP was significantly reduced (adjusted p-value < 0.05) to 22%, however, transfection efficiency for tdTomato was not significantly reduced (adjusted p-value > 0.05) compared to the single transgene conditions (Fig. 2e). When Turbofect was used as the cationic carrier to separately complex and deliver the two vectors, transfection efficiencies for EGFP and tdTomato were significantly reduced (adjusted p-value < 0.0001) to 38% and 33%, respectively, compared to the single transgene conditions (Fig. 2f). Finally, when two DNA vectors were delivered to D4 hBMSCs in separate complexes using Lipofectamine 3000, transfection efficiencies for EGFP and tdTomato were significantly reduced (adjusted p-value < 0.01) to 19% and 22%, respectively, compared to the single transgene conditions (Fig. 2g). However, when Turbofect was used as the cationic carrier to separately complex and deliver the two vectors, transfection efficiency for EGFP was not significantly reduced (adjusted p-value > 0.05) but was significantly reduced for tdTomato (adjusted p-value < 0.05) compared to the single transgene conditions (Fig. 2h).

When the two DNA vectors were delivered to D1 hAMSCs by first mixing the vectors together prior to formation of the cationic complexes (denoted as “[pE + pT]”, Fig. 1bii), transfection efficiencies were reduced (adjusted p-value < 0.05) to 24% for EGFP and 25% for tdTomato when Lipofectamine 3000 was used as the cationic carrier (Fig. 2a), but were not significantly reduced when Turbofect was used, compared to the single transgene conditions (Fig. 2b). When two DNA vectors were delivered to D2 hAMSCs in the same complex, transfection efficiencies for EGFP and tdTomato were not significantly reduced (adjusted p-value > 0.05) compared to the single transgene conditions when Lipofectamine 3000 was used as the cationic carrier (Fig. 2c). However, when Turbofect was used as the cationic carrier, transfection efficiencies for EGFP and tdTomato were significantly reduced (adjusted p-value < 0.001) to 24% and 25%, respectively, compared to the single transgene conditions (Fig. 2d). When two DNA vectors were delivered to D3 hBMSCs in the same complex, transfection efficiencies for EGFP and tdTomato were not significantly reduced (adjusted p-value > 0.05) compared to the single transgene conditions when Lipofectamine 3000 was used as the cationic carrier (Fig. 2e). However, when Turbofect was used as the cationic carrier, transfection efficiencies for EGFP and tdTomato were significantly reduced (adjusted p-value < 0.0001) to 41% for both transgenes when compared to the single transgene conditions (Fig. 2f). When two DNA vectors were delivered to D4 hBMSCs in the same complex, transfection efficiencies for EGFP and tdTomato were not significantly reduced (adjusted p-value > 0.05) when either Lipofectamine 3000 or Turbofect was used as the cationic carrier compared to single transgene conditions (Fig. 2g and h). Altogether, the data suggest that expression of two transgenes is less efficient than expression of a single transgene in hMSCs, however, delivering two single transgene DNA vectors in the same complex ([pE + pT]) is more efficient at expressing the individual transgenes than delivering two single transgene DNA vectors in different complexes ([pE]+[pT]).

We next determined the percentage of hMSCs that co-expressed EGFP and tdTomato following delivery of two DNA vectors each encoding for a single transgene, either complexed in separate (Fig. 1bi) or the same complexes (Fig. 1bii), by dividing the number of EGFP positive cells that were also tdTomato positive by the Hoescht count (i.e., cell count) in each well (see Assessment of Transfection Efficiency and Transgene Expression Levels in Materials and Methods section for more detail). The percentage of cells that co-expressed EGFP and tdTomato when two DNA vectors were delivered in separate complexes ([pE]+[pT]) was 11% when Lipofectamine 3000 was used as the cationic carrier and 23% when Turbofect as the cationic carrier was used for D1 hAMSCs (Fig. 2a and b). The percentage of cells that co-expressed EGFP and tdTomato in D2 hAMSCs when two DNA vectors were delivered in separate complexes was 7% when Lipofectamine 3000 was used as the cationic carrier and 13% when Turbofect was used (Fig. 2c and d). The percentage of cells that co-expressed EGFP and tdTomato in D3 hBMSCs when two DNA vectors were delivered in separate complexes was 14% when Lipofectamine 3000 was used as the cationic carrier and 26% when Turbofect was used (Fig. 2e and f). Lastly, the percentage of cells that co-expressed EGFP and tdTomato when two DNA vectors were delivered in separate complexes was 12% in D4 hBMSCs when Lipofectamine 3000 was used as the cationic carrier and 22% when Turbofect was used (Fig. 2g and h).

Finally, comparing the percentage of cells co-expressing EGFP and tdTomato when the two DNA vectors were delivered in the same complex to delivery of two DNA vectors in separate complexes showed a significant increase (adjusted p-value < 0.01) in co-expression transfection efficiencies regardless of the cationic carrier used or the donor of hMSC delivered to (Fig. 2a-g), except for D4 hBMSCs when Turbofect was used as the cationic carrier (Fig. 2h). However, the percentage of cells co-expressing EGFP and tdTomato in D4 hBMSCs when two DNA vectors were delivered in the same complex was still higher (34%) than delivery of two DNA vectors in separate complexes (22%; Fig. 2h), demonstrating that in all conditions studied with single transgene vectors, inclusion of both vectors within the same complex resulted in the highest co-expression transfection efficiencies.

Delivery of a single, bi-cistronic Vector in hMSCs

Next, we investigated whether delivering two transgenes on the same DNA vector, separated by an IRES or D2A sequence, could efficiently co-express both transgenes in hMSCs. Transfection efficiencies (i.e., number of EGFP, tdTomato, and co-expressing cells divided by total cell counts) were measured for bi-cistronic IRES and D2A DNA vector conditions and compared to single DNA vector conditions (i.e., delivery of a single DNA vector encoding either EGFP or tdTomato). It should be noted that in these studies, no expressionless plasmid was added to these conditions tested since a given mass of pDNA has similar moles of expression cassette, therefore, single transgene DNA vector transfection efficiencies were again measured to obtain a baseline using DNA doses that matched those of the bi-cistronic vectors (i.e., same mass of pDNA and moles of expression cassette). At this dose, single DNA vector transfection efficiencies for D1 hAMSCs were 35% and 32% for EGFP and tdTomato, respectively, when Lipofectamine 3000 was used as the cationic carrier, and 41% and 42%, respectively, when Turbofect was used as the cationic carrier (Fig. 3a and b). However, when the bi-cistronic IRES DNA vector was used at the same dose as the single transgene vectors, transfection efficiencies for EGFP and tdTomato were significantly reduced (adjusted p-value < 0.0001) to 29% and 18%, respectively, compared to the transfection efficiencies for the single transgene DNA vector conditions, but were not significantly reduced (adjusted p-value > 0.05) for the bi-cistronic D2A DNA vector, regardless of cationic carrier used (Fig. 3a and b). Furthermore, the bi-cistronic D2A DNA vector produced significantly higher transfection efficiencies (adjusted p-value < 0.0001) for both transgenes compared to the bi-cistronic IRES DNA vector, regardless of cationic carrier used (Fig. 3a and b). Single DNA vector transfection efficiencies for D2 hAMSCs were 22% and 24% for EGFP and tdTomato, respectively, when Lipofectamine 3000 was used as the cationic carrier, and 30% and 32%, respectively, when Turbofect was used as the cationic carrier (Fig. 3c and d). Transfection efficiencies for both EGFP and tdTomato were significantly reduced (adjusted p-value < 0.001) to 14% and 10% for the bi-cistronic IRES DNA vector compared to the transfection efficiencies for the single transgene DNA vector conditions, regardless of cationic carrier used (Fig. 3c and d). Transfection efficiencies for EGFP and tdTomato in D2 hAMSCs were not significantly reduced (adjusted p-value > 0.05) for the bi-cistronic D2A DNA vector compared to the single transgene conditions when Lipofectamine 3000 was used as the cationic carrier (Fig. 3c), but were significantly reduced (adjusted p-value < 0.05) to 16% and 22%, respectively, compared to the single transgene conditions when Turbofect was used as the cationic carrier (Fig. 3d). Conversely, the bi-cistronic D2A DNA vector produced significantly higher transfection efficiencies (adjusted p-value < 0.001) for EGFP and tdTomato compared to the bi-cistronic IRES DNA vector when Lipofectamine 3000 was used as the cationic carrier (Fig. 3c), while no significant difference in transfection efficiency for each transgene was seen between the bi-cistronic IRES and D2A DNA vectors when Turbofect was used as the cationic carrier (Fig. 3d).

Fig. 3

Delivery of Bi-Cistronic DNA Vectors with an IRES or D2A Sequence for Expression of Two Reporter Transgenes in hMSCs.  hMSCs were transfected with bi-cistronic DNA vectors complexed with Lipofectamine 3000 (a, c, e, & g) or Turbofect (b, d, f, & h) and transfection efficiencies for each transgene (EGFP, green bars, or tdTomato, red bars) were compared to the expression levels of single transgene vectors (pEGFP or ptdTomato) in D1 hAMSCs (a & b), D2 hAMSCs (c & d), D3 hBMSCs (e & f), and D4 hBMSCs (g & h). Transfection efficiencies for hMSCs expressing both EGFP and tdTomato (co-expression, yellow bars) was calculated by dividing the number of cells that were both EGFP and tdTomato positive by the total cell count (Hoechst, nuclear stain). All conditions have equal moles of expression cassette and mass of DNA delivered. Data represented as mean ± SEM (n = 6). * p < 0.05, **, p < 0.01, *** p < 0.001, **** p < 0.0001. ns, not significant (p > 0.05), as determined by a 2-way ANOVA with Tukey’s post hoc test

Single DNA vector transfection efficiencies for D3 hBMSCs were 41% and 39% for EGFP and tdTomato, respectively, when Lipofectamine 3000 was used as the cationic carrier (Fig. 3e), and 52% and 57%, respectively, when Turbofect was used as the cationic carrier (Fig. 3f). Transfection efficiencies for EGFP and tdTomato were significantly reduced (adjusted p-value < 0.0001) for the bi-cistronic IRES DNA vector compared to the transfection efficiencies for the single transgene DNA vector conditions, regardless of complexing reagent used, while the bi-cistronic D2A DNA vector only had a significant reduction in tdTomato transfection efficiency (adjusted p-value < 0.05) compared to the single transgene condition when Turbofect was used as the cationic carrier (Fig. 3e and f). Moreover, the bi-cistronic D2A DNA vector produced significantly higher transfection efficiencies (adjusted p-value < 0.0001) for EGFP and tdTomato compared to the bi-cistronic IRES DNA vector regardless of cationic carrier used (Fig. 3e and f). Lastly, single DNA vector transfection efficiencies for D4 hBMSCs were 33% and 33% for EGFP and tdTomato, respectively, when Lipofectamine 3000 was used as the cationic carrier, and 46% and 47%, respectively, when Turbofect was used as the cationic carrier (Fig. 3g and h). Transfection efficiencies for EGFP and tdTomato were significantly reduced (adjusted p-value < 0.05) for the bi-cistronic IRES DNA vector compared to the transfection efficiencies for the single transgene DNA vector conditions, regardless of cationic carrier used (Fig. 3g and h). Conversely, the transfection efficiencies for EGFP and tdTomato were not significantly reduced (adjusted p-value > 0.05) for the bi-cistrionic D2A DNA vector compared to the single transgene conditions, except for tdTomato when Turbofect was used as the cationic carrier (Fig. 3h). Furthermore, the bi-cistronic D2A DNA vector produced significantly higher transfection efficiencies (adjusted p-value < 0.001) for EGFP and tdTomato compared to the bi-cistronic IRES DNA vector regardless of cationic carrier used (Fig. 3g and h). Altogether, the data suggest that expressing two transgenes from a single, bi-cistronic DNA vector is less efficient than expressing the same transgene from a mono-cistronic DNA vector in hMSCs. Moreover, inclusion of a D2A peptide sequence between two distinct transgenes encoded on a single DNA vector can significantly increase expression of both transgenes compared to inclusion of an IRES between those transgenes.

Finally, when comparing the percentage of cells co-expressing both transgenes (i.e., number cells that express EGFP and tdTomato divided by total cell count) for the two bi-cistronic DNA vectors (Fig. 3a-h), the D2A DNA vector produced a significantly higher percentage of cells (adjusted p-value < 0.001) that expressed both EGFP and tdTomato compared to the IRES DNA vector in all donors regardless of cationic carrier used, except for D2 hAMSCs when Turbofect was used as the cationic carrier (Fig. 3d), suggesting that in all conditions studied with bi-cistronic vectors, inclusion of D2A peptide sequence resulted in the highest co-expression transfection efficiencies.

Identification of transfection strategies for simultaneous co-expression of two transgenes in hMSCs

Next, we calculated the percentage of cells that were successfully transfected (i.e., expressing EGFP) that were simultaneously expressing both transgenes from the bi-cistronic IRES and D2A DNA vector conditions by dividing the number of EGFP positive cells that were also tdTomato positive by the number of EGFP positive cells (Fig. 4). The reasoning behind normalizing to EGFP positive cells, as well as counting EGFP positive cells that are also tdTomato positive, as opposed to normalizing to tdTomato positive cells or counting tdTomato positive cells that are also EGFP positive, is that for D2A sequences, translation of the upstream transgene (EGFP) is required before translation of the downstream transgene (tdTomato) can occur. The bi-cistonic D2A DNA vector produced significantly more successfully transfected cells (adjusted p-value < 0.01) that were co-expressing EGFP and tdTomato compared to the bi-cistronic IRES DNA vector regardless of hMSC donor, tissue source, or cationic carrier used (Fig. 4). The bi-cistronic IRES DNA vector resulted in 60–84% of successfully EGFP transfected cells simultaneously co-expressing tdTomato, whereas the bi-cistronic D2A DNA vector resulted in 99-100% of successfully EGFP transfected cells simultaneously co-expressing tdTomato, regardless of hMSC donor, tissue source, or cationic carrier used (Fig. 4, Supplemental Fig. 1), suggesting that in all conditions studied with bi-cistronic vectors, inclusion of D2A peptide sequence resulted in the highest efficiency of expressing the downstream transgene (tdTomato) in cells that were expressing the upstream transgene (EGFP).

Fig. 4

Percent of Transfected hMSCs Simultaneously Expressing Both Transgenes from Bi-Cistronic DNA Vectors. hMSCs were transfected with either the bi-cistronic pD2A or pIRES DNA vector complexed with Lipofectamine 3000 (a, c, e, & g) or Turbofect (b, d, f, & h) and assayed for transgene expression 24 h after transfection for D1 hAMSCs (a & b), D2 hAMSCs (c & d), D3 hBMSCs (e & f), and D4 hBMSCs (g & h). The percent of EGFP expressing cells that are expressing tdTomato was calculated by dividing the number of EGFP positive cells that were also tdTomato positive by the number of EGFP positive cells. The reasoning behind normalizing to EGFP positive cells, as well as counting EGFP positive cells that are also tdTomato positive, as opposed to normalizing to tdTomato positive cells or tdTomato positive cells that are also EGFP positive, is that translation of the upstream transgene (EGFP) is required for D2A sequences before translation of the downstream transgene can occur. Data represented as mean ± SEM (n = 6). * indicates significance relative to pIRES. *** p < 0.001, **** p < 0.0001, as determined by a two-tailed T-test

We next directly compared delivery strategies for expressing two transgenes (i.e., [pE]+[pT], [pE + pT], or delivery of bi-cistronic DNA vectors that express two transgenes via an IRES or D2A sequence; Fig. 5). To properly compare these conditions, addition of expressionless vector to the bi-cistronic DNA vector conditions was needed to equalize mass of pDNA and molarity of each transgene for each condition. Directly comparing delivery of two DNA vectors in separate or the same complex, to delivery of a single bi-cistronic DNA vector with an IRES or D2A with both mass of DNA and molarity of expression cassette normalized across all conditions, showed that inclusion of an IRES sequence resulted in significantly reduced (adjusted p-value < 0.05) expression of tdTomato compared to all other conditions (Fig. 5a-h). However, the inclusion of an IRES sequence did not significantly reduce tdTomato expression compared to delivery of two DNA vectors in separate complexes, for all donors using Lipofectamine 3000 as the cationic carrier (adjusted p-value > 0.05). Furthermore, the bi-cistronic D2A DNA vector led to significantly higher transfection efficiencies (adjusted p-value < 0.05) of cells co-expressing EGFP and tdTomato compared to all other conditions, except when compared to delivery of two transgenes in the same complex. However, the bi-cistronic D2A DNA vector showed significantly higher co-expression compared to delivery of two transgene in the same complex in D3 hBMSCs when Turbofect was the used as the cationic carrier (Fig. 5f). These results suggests that in all conditions studied, delivery of a single DNA vector with a D2A peptide sequence, as well as delivery of two DNA vectors in the same complex, resulted in the highest co-expression transfection efficiencies.

Fig. 5

Equal Mass of DNA and Copy Number of Transgenes Comparison for Delivery of Two DNA Vectors to Delivery of Bi-Cistronic DNA Vectors for Expression of Two Transgenes in hMSCs.  hMSCs were transfected with single transgene DNA vectors delivered in separate complexes ([pE]+[pT]) the same complex ([pE + pT]) or bi-cistronic DNA vectors (pD2A or pIRES) with an equal mass of DNA and copy number of transgenes delivered. All four conditions were complexed with Lipofectamine 3000 (a, c, e, & f) or Turbofect (b, d, f, & h). Transgene expression for each condition was assayed 24 h after transfection for D1 hAMSCs (a & b), D2 hAMSCs (c & d), D3 hBMSCs (e & f), and D4 hBMSCs (g & h). Expression of both transgenes simultaneously in hMSCs (co-expression) was calculated by dividing the number of cells that were both EGFP and tdTomato positive by the total cell count (Hoechst, nuclear stain). Data represented as mean ± SEM (n = 6). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. ns, not significant (p > 0.05), as determined by a 2-way ANOVA with Tukey’s post hoc test

Finally, we calculated the percentage of EGFP-expressing cells that were simultaneously expressing Tdtomato for delivery of two DNA vectors in separate or the same complex, as well as for the bi-cistronic IRES and D2A DNA vectors, with mass of DNA and moles of each transgene delivered equalized across all conditions. Delivery of two DNA vectors in separate complexes led to 56–73% of successfully EGFP-transfected cells also expressing tdTomato, across all donors and cationic carries tested (Fig. 6). Similarly, the bi-cistronic IRES DNA vector led to 58–80% of successfully EGFP-transfected cells simultaneously expressing tdTomato across all donors and cationic carries tested (Fig. 6). Conversely, delivery of two DNA vectors in the same complex led to 90–96% of successfully EGFP-transfected cells simultaneously expressing Tdtomato across all donors and cationic carriers tested, which was significantly higher (adjusted p-value < 0.01) than delivery of two DNA vectors in separate complexes and the bi-cistronic IRES DNA vector for all donors and cationic carriers tested (Fig. 6). However, the bi-cistronic D2A DNA vector led to 99–100% of successfully EGFP-expressing cells simultaneously expressing Tdtomato across all donors and cationic carriers tested, which was significantly higher (adjusted p-value < 0.05) than all tested conditions except for delivery of two DNA vectors in the same complex using Lipofectamine 3000 as the cationic carrier in D2, D3, and D4 hMSCs and when using Turbofect as the cationic carrier in D2 hAMSCs (Fig. 6). Altogether, these results suggest that a bi-cistronic DNA vector with a D2A peptide sequence can mediate up to 100% of successfully transfected hMSCs (i.e., cells expressing the first transgene) simultaneously expressing the second transgene, and delivery of two DNA vectors in the same complex can mediate up to 96% of successfully transfected hMSCs simultaneously expressing the second transgene in multiple donors, regardless of cationic carrier.

Fig. 6

Comparison of Percent of Transfected hMSCs Simultaneously Expressing Both Transgenes from Delivery of Two DNA Vectors to Delivery of Bi-Cistronic DNA Vectors when Equal Mass of DNA and Copy Number of Transgenes are Delivered.  hMSCs were transfected with single transgene DNA vectors delivered in separate complexes ([pE]+[pT]), the same complex ([pE + pT]) or bi-cistronic DNA vectors (pD2A or pIRES) with an equal mass of DNA and copy number of transgenes delivered. All four conditions were complexed with Lipofectamine 3000 (a, c, e, & g) or Turbofect (b, d, f, & h). Transgene expression for each condition was assayed 24 h after transfection for D1 hAMSCs (a & b), D2 hAMSCs (c & d), D3 hBMSCs (e & f), and D4 hBMSCs (g & h). The percent of EGFP expressing cells that are expressing tdTomato was calculated by dividing the number of EGFP positive cells that were also tdTomato positive by the number of EGFP positive cells. The reasoning behind normalizing to EGFP positive cells, as well as counting EGFP positive cells that are also tdTomato positive, as opposed to normalizing to tdTomato positive cells or tdTomato positive cells that are also EGFP positive, is that translation of the upstream transgene (EGFP) is required for D2A sequences before translation of the downstream transgene can occur. Data represented as mean ± SEM (n = 6). * indicates significance relative to [pE]+[pT]. *** p < 0.001, **** p < 0.0001. # indicates significance relative to [pE+pT]. # p < 0.05, ## p < 0.01. $ indicates significance relative to pIRES. $ $ p < 0.01, $ $ $ p < 0.001, $ $ $ $ p < 0.0001, as determined by a 2-way ANOVA with Tukey’s post hoc test. Only significant comparisons are shown for clarity