Tracking Plasmid Delivery Using a Fluorescent Plasmid in T cells
The relatively low transfection efficiency that was obtained with
Lipofectamine in the primary T cells could be due to multiple different
steps in the transfection process, including cellular uptake, nuclear
translocation, transcription, and translation of the transgene. However,
since other groups have achieved high transfection efficiencies (up to
70%) with electroporation of similar plasmids and synthetic mRNAs in
primary T cells,24,46–48 we hypothesized that
cellular uptake may be the key limiting step for Lipofectamine (instead
of nuclear uptake or transcription/translation). To test this
hypothesis, we used a fluorescein-labeled plasmid (Mirus Bio™, MIR 7906)
to track the delivery of fluorescent plasmid DNA to the cells with flow
cytometry and confocal microscopy.
First of all, as shown in Figure 5A, transfection of adherent PC-3 cells
(a cell line that is relatively permissive to transfection) with
Lipofectamine and pEF-GFP provided a very high percentage of
EGFP+ cells (91±1.1% EGFP+).
Likewise, a similarly high percentage of the PC-3 cells fluoresced when
transfected with the fluorescein-labeled plasmid (96±0.94%
fluorescein+), demonstrating that the plasmid was
successfully delivered to almost all the PC-3 cells. Interestingly, the
transfection efficiency observed with pEF-GFP for the Jurkat cells
(51±9.7% EGFP+) was more modest, but the percentage
of fluorescein+ cells (74±4.5%) was significantly higher. The mean
fluorescence (Figure 5B) of the PC-3 cells was also significantly higher
than the mean fluorescence of the Jurkat cells in both types of
transfections.
Similar trends were also observed with the primary T cells (Figure
5A/B). The percentage of fluorescein+ primary T cells
and the mean fluorescence of the fluorescein+ primary
T cells was significantly lower than both Jurkat T cells and the PC-3
cells. There was also a more dramatic disparity between the fraction of
fluorescein+ primary T cells (51±11.2%
EGFP+) and the percentage of primary T cells
expressing EGFP (8.1±0.8% EGFP+). This stark
difference suggests that the lipoplexes bind to a large fraction of the
primary T cells, but many of those lipoplexes may fail to either enter
the cell via endocytosis, escape the endosome into the cytosol, or enter
the nucleus.
To determine if the fluorescent lipoplexes were bound to the surface of
the cells or taken up into the cytoplasm or nucleus, cells transfected
with the fluorescein-labeled plasmid were treated with trypsin-EDTA for
up to 30 minutes to disrupt any potential interactions between cell
surface proteins on the cell surface and the lipoplex (Figure 5C/D).
Hypothetically, if the fluorescent lipoplexes were simply bound to the
cell membrane, then this trypsinization would detach them, leading to a
significantly lower mean fluorescence and percentage of
fluorescein+ cells. Indeed, trypsinization had no
significant effect on the percentage of fluorescein+PC-3 cells or their mean fluorescence, suggesting that most of the
fluorescent lipoplexes were in the cytosol or nucleus of the PC-3 cells.
In contrast, trypsinization significantly decreased the mean
fluorescence and the percentage of fluorescein+ Jurkat
cells to a level that was comparable to the percentage of
EGFP+ cells shown in Figure 5A. A similar, although
not statistically significant, decrease in the percentage of
fluorescein+ cells was also observed in the primary T
cells. However, the mean fluorescence of the primary T cells
significantly decreased and was virtually eliminated after 30 minutes of
trypsinization (Figure 5D). Altogether, these findings suggest that
cellular uptake may be a limiting step for gene delivery in both Jurkat
and primary T cells.
To further investigate the localization of the fluorescein-labeled
plasmid in the PC-3 and primary T cells, fluorescent microscopy was used
to visualize the fluorescent lipoplexes after transfection (Figure 6).
In these experiments, cells were also stained with Hoescht 33342 nuclear
stain and the Biotium CellBrite red cytoplasmic membrane dye to
visualize the nucleus and cytoplasm, respectively. Overall, similar
trends in fluorescence were observed across the cell lines. A widespread
distribution of bright EGFP fluorescence within the cytoplasm was
observed in PC-3 cells transfected with pEF-EGFP (indicating successful
transgene expression), while cells transfected with the
fluorescein-labeled plasmid exhibited small regions of concentrated
fluorescence inside the cytoplasm and nucleus that indicated successful
plasmid uptake and nuclear delivery (Figure 6A). In addition, z-stacking
with 0.5 µm slices from the top edge to the bottom edge of the PC-3
cells also confirmed that the fluorescent lipoplexes were inside the
cell (individual images areshown in Figure 6C, while compiled z-stacks
are shown in a video in Figure S5).
In contrast, transfection of CD3+ primary T cells with
pEF-EGFP only yielded a small fraction (3.9±0.24%) of
EGFP+ cells and the primary T cells that did fluoresce
were much dimmer than the PC-3 cells (in agreement with the flow
cytometry data shown in Figure 5B). Confocal microscopy z-stacking
images also showed that the fluorescent lipoplexes appeared to be
localized to the outside of the cell membrane instead of being
internalized by the cell (Figures 6C and S6). These observations further
support the notion that lipoplexes may successfully bind to primary T
cells, but endocytosis of the lipoplexes appears to be limited.
Transcriptome
Analysis of Primary T Cells
The experiments with the fluorescein-labeled plasmid (Figures 5-6)
seemed to suggest that impaired cellular uptake of the plasmid may be
partially responsible for the T cells’ resistance to transfection, but
the observation that 30-40% of the primary T cells remained
fluorescein+ after trypsinization suggested that other
intracellular mechanisms may also impair endosomal escape, nuclear
delivery, and/or expression of the transgene. In an effort to detect any
additional mechanisms that could inhibit transgene delivery or
expression and provide a rationale for the relatively low transfection
efficiency of primary T cells, mRNA sequencing was used to investigate
the transcriptomes of PC-3, Jurkat, and primary T cells in the absence
or presence of lipoplexes. These experiments were motivated by previous
studies that showed both viral vectors and Lipofectamine can activate
innate immune response pathways which trigger the expression of
anti-viral genes (TLRs, MyD88, IRFs) that can hinder transduction in
some cell types.54
The complete mRNA-seq data (fastq and bam files, along with a
spreadsheet of all FPKM values) from these experiments are available at
the NCBI GEO repository (GEO Accession# GSE151759). Overall, one of the
most significant differences that was observed in the gene expression
profiles of the PC-3 cells and both types of T cells (Table 1) was the
absence of several heparan sulfate proteoglycans (HSPGs), which are also
known as syndecans (SDCs). As shown in Table 1, HSPG2 and all the
sydecans were expressed in PC-3 cells, but mostly absent in both Jurkat
and primary T cells (with and without lipoplexes). SDC3 was the only
syndecan expressed in Jurkat cells, but its role in endocytosis and gene
delivery has not yet been established. SDC4 was expressed in primary T
cells (as expected for activated T cells), but SDC2 was not detected.
Similar results were reported by one study that showed 100-fold lower
levels of HSPG expression in Jurkat T cells compared to HeLa
cells.77 Low levels of HSPG expression have also been
observed in primary T cells, but T cell activation can upregulate SDC2
and SDC4 expression.58,78
This general lack of syndecan expression may explain the significantly
lower transfection efficiencies shown in Figure 5A for the Jurkat and
primary T cells relative to the PC-3 cells (which express HSPG2 and all
the syndecans). Indeed, while HSPGs are best known for their roles in
the attachment of adherent cells to the extracellular matrix or tissue
culture plates, they are also directly involved in gene delivery, since
they regulate endocytosis and their negatively charged sulfate groups
are involved in the initial binding of several viruses and positively
charged polyplexes or lipoplexes.79 Indeed,
overexpression of SDC1, SDC2, and SDC4 enhances the transfection
efficiency of liposomes in K562 cells, although overexpression of SDC2
has also been shown to inhibit PEI-mediated gene
delivery.67,80 Alternatively, blocking the sulfation
of SDCs has also been shown to inhibit endocytosis and gene
delivery.57,67
As previously mentioned, mRNA-sequencing was performed in the different
cell lines both in the absence and presence of lipoplexes. The rationale
for comparing transfected and untransfected cells was to determine if
there were any host cell genes that were upregulated in response to
transfection of double-stranded plasmid DNA. Indeed, there are many
examples of genes that are induced or upregulated when dsDNA is detected
in the cytoplasm and many of these upregulated genes have potent
anti-viral functions that are designed to inhibit the replication of
viruses.68,69 Unfortunately, many of these genes can
also inhibit non-viral transgene delivery or
expression.70
Transfection of PC-3 cells led to the induction or upregulation of
hundreds of cytokines and cytokine-stimulated genes (CSGs, data not
shown), some of which can inhibit transgene delivery (e.g., IFITM 1, 2,
and 3) or translation (e.g., IFIT 1 and 2). Nonetheless, the
transfection efficiency and mean GFP levels were still high in PC-3
cells, suggesting that many of these upregulated genes may be
inconsequential for non-viral pDNA or transgene delivery and expression
in PC-3 cells.
In contrast to PC-3 cells, only a few differentially expressed genes
were significantly upregulated at least 3-fold in the Jurkat and primary
T cells (Table 2). This result is somewhat unanticipated, since the
requisite cytosolic DNA sensors (e.g., cGAS and IFI16) and the
downstream effectors (e.g., STING, TBK1, and IRFs) that are necessary to
detect foreign DNA and induce the expression of cytokines and CSGs were
detected in the Jurkat and primary T cells (data not shown). This lack
of an innate immune response to dsDNA has been previously reported by
other groups, suggesting that T cells may lack an unknown component of
the DNA sensing pathways or they somehow repress
CSGs.8 Nonetheless, some metallothioneins were
significantly upregulated in the Jurkat (MT1F, MT2A) and primary T cell
lines (MT1H) after transfection. Metallothioneins are involved mainly in
metal binding, often to zinc, but they have also been implicated in
immune regulation and the response to bacterial and viral
infections.60,61 However, the role of metallothioneins
in transgene delivery and expression has not yet been determined.
Although T cells did not significantly upregulate cytokines or CSGs in
response to transfection, one interesting trend that emerged when
comparing the transcriptomes of the T cells to the PC-3 cells was that
multiple CSGs that were upregulated in PC-3 cells after transfection
were constitutively expressed at relatively high levels in the T cell
lines (Table 2). For example, each member of the pyrin and HIN (PYHIN)
domain DNA sensor family (IFI16, AIM2, and PYHIN1/IFIX) was detected in
both untransfected and transfected primary T cells. In contrast, AIM2
was only expressed in PC-3 cells after transfection and IFI16 was highly
upregulated in transfected PC-3 cells, suggesting that these DNA sensors
play an important role in the innate immune response to dsDNA in PC-3
cells. This is a particularly intriguing observation, because after AIM2
and IFI16 bind dsDNA in the cytoplasm, they form an inflammasome complex
with PYCARD, Caspase 1/8, and Gasdermin D (all of which were expressed
at detectable levels in primary T cells, but not Jurkat cells) that can
induce inflammation, apoptosis, and pyroptosis.64Pyroptosis has also been observed in primary T cells during abortive HIV
infection, in which double-stranded cDNA is generated in the cytoplasm
by the virus.63 Therefore, while the AIM2 and IFI16
inflammasome pathways must be induced (AIM2 & CASP1) or upregulated in
PC-3 cells, it appears that primary T cells constitutively express the
genes in these pathways, which may lead to higher levels of apoptosis
upon transfection with plasmid DNA and the decrease in proliferation
shown in Figure 1. Indeed, several other studies have reported that
dsDNA is highly toxic in T cells.71,72