【专题讨论】外文文献讨论（一）---------欢迎大家积极参与、共同提高
丁香园论坛
3919
本人的想法：拿出自己研究过的文献（外文），供大家提出问题，讨论、翻译。相信大家的讨论一定会碰撞出新的ideas来。（已经搜索了相关内容，应该不是重复法帖，请班主支持！）
相互交流、享受提高！！！
Small Interfering RNA-Mediated Gene Silencing in T
Lymphocytes1
Michael T. McManus, Brian B. Haines, Christopher P. Dillon, Charles E. Whitehurst,
Luk van Parijs, Jianzhu Chen, and Phillip A. Sharp2
Introduction of small interfering RNAs (siRNAs) into a cell can cause a specific interference of gene expression known as RNA
interference (RNAi). However, RNAi activity in lymphocytes and in normal primary mammalian cells has not been thoroughly
demonstrated. In this report, we show that siRNAs complementary to CD4 and CD8 specifically reduce surface expression of
these coreceptors and their respective mRNA in a thymoma cell line model. We show that RNAi activity is only caused by a subset
of siRNAs complementary to the mRNA target and that ineffective siRNAs can compete with effective siRNAs. Using primary
differentiated T lymphocytes, we provide the first evidence of siRNA-mediated RNAi gene silencing in normal nontransformed
somatic mammalian lymphocytes. The Journal of Immunology, 2002, 169: 5754–5760.
I ntroduction of dsRNA into an organism can cause specific
interference of gene expression (1). This phenomenon,
known as RNA interference (RNAi),3 results from a specific
targeting of mRNA for degradation by an incompletely characterized
cellular machinery present in plant, invertebrate, and mammalian
cells (2, 3). The proteins mediating RNAi are part of an
evolutionarily conserved cellular pathway that processes endogenous
cellular RNAs to silence developmentally important genes (4,
5). In RNAi, the protein Dicer, an RNase III enzyme, is probably
responsible for the processing of dsRNA into short interfering
RNA (siRNA). Functional screens conducted in plants and worms
have identified a number of other conserved genes participating in
the RNAi pathway. These genes include a number of different
helicases, a RNA-dependent RNA polymerase, an exonuclease,
dsRNA-binding proteins, and novel genes of unknown function
(for recent reviews, Refs. 6, 7, 8, 9, and 10).
Mammalian RNAi was first described in mouse embryos using
long dsRNA (11, 12). Then, following the analysis of the structure
of the intermediate in this process, small interfering RNAs
(siRNAs) were used to silence genes in mammalian tissue culture
(13, 14). Most of the RNAi pathway genes discovered in plant and
worm screens are also present in mouse and human sequence databases,
supporting evidence that a conserved RNAi pathway exists
in mammals. One of the more notable exceptions is the RNAdependent
RNA polymerase gene, which has been shown to beinvolved in the amplification of the dsRNA in Caenorhabditis elegans
(15, 16). This might imply that perpetuation of the RNAi
response in mammals differs from that of lower organisms.
Recent reports have demonstrated gene silencing by siRNA in
mammalian cells (17–22). However, despite these initial reports,
many uncertainties remain concerning the mechanism, physiologic
relevance, and ubiquity of RNAi in mammalian cells. Although
studies in tumor cell lines have demonstrated siRNA-mediated
RNAi, it remains a major question as to whether primary cells
from fresh tissues can undergo the RNAi response. Furthermore,
little is known about the efficiency and longevity of siRNA-mediated
RNAi gene suppression. In this report, we provide fundamental
insight into the siRNA-mediated RNAi mechanism using a thymoma-
derived cell line model to demonstrate for the first time the
occurrence of RNAi in primary T lymphocytes.
Materials and Methods
Cell culture
E10 is an immature double-positive thymocyte line derived from a TCR-
and p53 double-mutant mouse of a mixed 129/Sv C57BL/6 background
as described (23). These cells, which proliferated vigorously, were maintained
at a maximal concentration of 2 106 cells/ml and were propagated
in complete medium: DMEM supplemented with 10% heat-inactivated
FCS, 2 mM L-glutamine, 100 U/ml penicillin, 100 g/ml streptomycin, and
50 M 2-ME. Cell culture of primary lymphocytes: cells from the spleen
and lymph nodes of DO11.10 TCR-transgenic mice (a generous gift from
Dr. C. London, University of California, Davis, CA) were activated for 3
days with 1 g/ml OVA peptide (residues 323–339) in RPMI medium
containing 10% FBS.
Transfection
For electroporations, 2.5 mol dsRNA and/or 20 g of pEGFP-N3 plasmid
(Clontech Laboratories, Palo Alto, CA) were added to prechilled 0.4-cm
electrode gap cuvettes (Bio-Rad, Hercules, CA). E10 cells (1.5 107)
were resuspended to 3 107 cells/ml in cold serum-free RPMI, added to
the cuvettes, mixed, and pulsed once at 300 mV, 975 F with a Gene
Pulser electroporator II (Bio-Rad). Cells were plated into 6-well culture
plates containing 8 ml of complete medium and were incubated at 37°C in
a humidified 5% CO2 chamber. Cell viability immediately after electroporation
was typically around 60%. For cationic lipid transfections, 2 g of
plasmid DNA and 100 nmol siRNAs were used per 106 cells, and transfection
followed manufacturer’s recommended protocol. Transfection of
primary lymphocytes: activated DO11.10 T cells were electroporated as
above, except that the cells were resuspended to 6 107 cells/ml in cold
serum-free RPMI and the pulse voltage was 310 mV. After electroporation,
the cells were put into four wells of a 24-well plate, each containing 1 mlof RPMI supplemented with 1 ng/ml IL-2 (BioSource International, Camarillo,
CA). siRNA oligos (Dharmacon, Lafayette, CO) used were as follows
(sense strand is given): effective CD4 siRNA, CD4 no. 4, (sense)
gagccauaaucucaucugadgdg, (anti-sense) ucagaugagauuauggcucdtdt; effective
CD8 siRNA, CD8 no. 4, (sense) gcuacaacuacuacaugacdtdt, (antisense)
gucauguaguaguuguagcdtdt; ineffective siRNAs, CD8 no. 1, (sense) gaaaa
uggacgccgaacuudgdg, (anti-sense), aaguucggcguccauuuucdtdt; CD8 no. 2,
(sense) cgugggacgagaagcugaadtdt, (antisense) uucagcuucucgucccacgdtdt;
CD8 no. 3 (sense) aauuguguaaaauggcaccgcdcda, (antisense) ggcggugc
cauuuuacacaadtdt; CD4 no. 1, (sense) ggagaccaccaugugccgadgdc, (antisense)
ucggcacaugguggucuccdtdt; CD4 no. 2, (sense) ggcagagaaggauucu
uucdtdt, (anti-sense) gaaagaauccuucucugccdtdt; CD4 no. 3, (sense)
ccaccugcguccugucucadtdc, (antisense) gugguggacgcaggacagadgdt; CD4 no. 5
(sense) ccaccugcguccugucucadtdc, (antisense) ucagaugagauuauggcucdtdt.
Flow cytometry
E10 cells (1 106) were washed once in FACS buffer (PBS supplemented
with 2% FCS and 0.01% sodium azide), resuspended to 100 l, and
stained directly with PE-conjugated anti-CD4 (clone RM4-5) or allophycocyanin-
conjugated anti-CD8 mAbs, and in some experiments with PEor
allophycocyanin-conjugated anti-mouse Thy-1.2 (clone 53-2.1) mAb.
All mAbs were from BD PharMingen (San Diego, CA). The stained cells
were washed once, then resuspended in 200 l FACS buffer containing 200
ng/ml propidium iodide (PI). Unstained and singly stained controls were
included in every experiment. 3A9, a T cell hybridoma line that had been
infected with a MIGW green fluorescent protein (GFP) retrovirus was included
when GFP expression was analyzed. Cell data were collected on a
FACSCalibur flow cytometer (BD Biosciences, San Jose, CA) and fourcolor
analyses (GFP, PE, PI, and allophycocyanin) were done with
CellQuest software (BD Biosciences). All data were collected by analyses
performed on 1 104 PI-negative events (viable cells). For the primary T
cell studies, activated cells were analyzed as above, except that allophycocyanin-
conjugated anti-CD4 and PE-conjugated anti-CD8 were used,
and 5 104 PI-negative events were analyzed.
Northern blot analysis of mRNA
Cells were lysed in TRIzol reagent (Life Technologies, Grand Island, NY)
and total cellular RNA was purified according to manufacturer’s instructions.
RNA (10 g) was fractionated on a denaturing 1% formaldehyde/
agarose gel and transferred to a nitrocellulose membrane. Blots were hybridized
overnight with 32P-labeled CD4 (818 bp) or CD8 (596 bp) cDNA
fragments. After washes, blots were analyzed by a PhosphorImager (Molecular
Dynamics).
Results
siRNAs transiently induce silencing in murine thymocyte
cell lines
To study RNAi, siRNAs are typically delivered into cells by carrier-
mediated transfection reagents. We developed an experimental
system using a thymoma-derived cell line, E10 (23), wherein we
studied the use of siRNAs to silence either CD4 or CD8, using
the other marker as an internal specificity control. However, typical
of lymphocytes, E10 is insensitive to several different cationic
and noncationic transfection reagents and thus electroporation was
used to introduce siRNAs. Using this method, 20% of the cell
population expressed GFP from a transfected reporter vector.
When CD4 or CD8 siRNAs were electroporated into E10, a
marked reduction in surface CD4 or CD8 expression, respectively,
occurred 36 h later. Flow cytometry analysis showed that
most of the cells were transfected and expression levels were reduced
5-fold below wild-type expression levels (Fig. 1A). The
degree of reduction of CD8 was frequently more pronounced
than that of CD4 and, in both cases, a small population of cells
appeared to be either untransfected or not responsive to the siRNA
treatment. In repeated experiments, typically 70–95% of the cells
exhibited a 5-fold reduction in CD8 expression, although
sometimes a smaller fraction of cells down-regulated CD8 to a
greater degree (Fig. 1A).
Elbashir et al. (13) reported that RNAi-induced silencing could
be maintained for 2 wk in HeLa cells, although neither the extent
of silencing nor the number of cell divisions was reported. A timecourse assay was performed in CD8 siRNA-transfected E10
cells. GFP was included in these transfections to investigate the
relationship between the uptake and expression of plasmid DNA
and siRNAs. Because these experiments were transient transfections,
cell doubling results in a decrease in GFP fluorescence intensity
and number of GFP-positive cells (Fig. 1B, NS RNA).
When CD8 siRNAs were cotransfected with the GFP reporter
vector, CD8 expression, but not GFP expression, was markedly
reduced (see Fig. 1B, 24 h). Several cell populations were evident,
with the major CD8 silenced population displaying 5-fold reduced
CD8 expression. The majority of cells within this population
did not express GFP. However, cells that did express GFP
also silenced CD8. This corresponded to 20% of the total cells,
similar to the control GFP alone (Fig. 1B, NS siRNA, 24 h). This
indicates that all of the cells expressing GFP also received an adequate
level of siRNAs to silence CD8. In addition, a large fraction
of cells incorporated biologically active levels of siRNAs andyet did not express plasmid DNA. In this experiment, time points
were taken over a period of 6 days. At each time point, one-half of
the cells were removed from the dish and replaced with fresh medium.
The collected cells were stained for CD8 and analyzed by
flow cytometry (Fig. 1). A decrease in CD8 surface expression
was detectable at 12 h posttransfection, with maximal silencing at
36 h. By 96 h, nearly all of the cells expressed wild-type levels of
CD8. Thus, the RNAi effect in these T cells is a transient
phenomena.
In these experiments, there was a dramatic decrease in GFP
expression over time, which was likely a result of dilution of the
plasmid or potentially due to toxicity of high GFP expression.
Because 100% of the GFP-expressing cells exhibited CD8 silencing,
it was possible to monitor the “fate” of this subset of
silenced cells. The T cells that actively underwent CD8 silencing
continued to express GFP over the time course, to the same level
as the control population of cells that were not transfected with
siRNAs (compare nonspecific RNA to CD8 siRNA). At 96 h,
5% of the total cells were GFP-positive in cells treated with
nonspecific siRNAs and in CD8 siRNA-treated samples. These
few remaining GFP-positive cells exhibited normal levels of
CD8 expression. This suggests that the cells did not specifically
undergo apoptosis as a result of siRNA transfection and subsequent
CD8 silencing.
Specificity of siRNA-mediated silencing
Although the GFP transgene expression was not affected during
CD8 silencing, the expression of endogenous genes might have
been nonspecifically affected. To address this question, the expression
levels of CD4 and Thy1.2 T cell markers were examined in
cells actively undergoing CD8 silencing. Examination of these
markers revealed that there was no reduction of nontargeted gene
expression when compared with the control nontransfected cells
(Fig. 2A), even over extended times (not shown). Although unlikely
for this cell line, an additional analysis confirmed that the T
cells did not become activated, as they do not up-regulate CD69
(Fig. 2A). Together, these experiments confirm the specificity of
siRNA-mediated CD8 silencing.
Stability of targeted CD8 mRNA
Short temporal RNAs such as lin-4 and let-7 mediate silencing by
binding to the 3-untranslated region (UTR), thus suppressing
translation (24–26). This is in marked contrast to the posttranscriptional
mRNA degradation effected by siRNAs. To distinguish between
these two potential mechanisms for CD8 silencing, a time
course Northern blot analysis of CD8 mRNA was performed.
The process of silencing did not appreciably affect the growth rate,
as compared with control nonspecific siRNA transfections performed
in parallel (not shown). Flow cytometry analysis indicated
that the RNAi response in these cells lasted 3–4 days (8–10 cell
doublings), which corresponds to an 100-fold increase in cell
mass (Fig. 2B). Time course analysis was performed in four independent
experiments and expression of CD8 was typically suppressed
5-fold or greater.
At various time points, a fraction of the cells was used to isolate
total RNA for Northern blot analysis (Fig. 2C). The CD8 mRNA
was resolved into two bands, due to alternative splicing (27, 28).
Levels of CD8 mRNA decreased during the course of CD8
silencing. Densitometric analysis of the CD8 mRNA bands was
performed and normalized to the internal control CD4 band. At the
point of maximal silencing, mRNA levels decrease only 2.5-fold.
This value is not commensurate with the 5-fold decrease in protein
expression determined by the flow cytometric analysis. However,
this RNA was prepared from total cells in which 30% of the
cells did not exhibit any silencing. When corrected for this reduction,
CD8 mRNA was nearly proportionate to levels in reduction
of CD8 protein. These Northern blots were performed multiple
times with similar results. Thus, although it is clear that CD8
mRNA decreases, we cannot rule out additional silencing phenomena
such as cotranslational repression.
Regional sensitivity of an mRNA to silencing by a siRNA
A major outstanding question is whether any region of a mRNA
can serve as an effective target for siRNA-directed silencing. Several
different siRNAs that targeted different regions of the CD8阿尔法mRNA were tested. Of the first two CD8 siRNAs that were transfected,
only one was active. To more quantitatively examine this
difference, cells were transfected with varying amounts of siRNAs
and CD8 expression was measured by flow cytometry. Cells undergoing
silencing were quantified and compared with control
nonspecific siRNA treatment (Fig. 3A). For the effective CD8
siRNA, picomolar amounts were sufficient to induce some silencing
and higher amounts produced a graded response. For the noneffective
CD8 siRNA, even at the highest concentration tested,
there was no activity.
As these studies progressed, we observed that the majority of the
synthetic CD4 and CD8 siRNAs were noneffective at silencing.
For CD8, four different siRNAs were synthesized and tested in
the flow cytometry assay: one overlapped the start codon, one
which targeted the open reading frame (ORF), one which overlapped
the stop codon, and one which targeted the 3-UTR 15 nt
after the stop codon. Only the siRNA which targeted the 3-UTR
15 nt after the stop codon effectively silenced CD8 expression.
For CD4, five siRNAs were synthesized which targeted corresponding
regions to those for the CD8 mRNA (Fig. 3B). In this
case, only the siRNA that targeted the stop codon was effective at
reducing CD4 expression levels. An examination of the nucleotide
sequences did not reveal any obvious differences between the effective
and ineffective siRNAs.
For each of the above siRNAs, the silencing assay was performed
at different siRNA concentrations. None of the inactive
siRNAs generated detectable silencing at five times the highest
concentration of the active siRNAs (Fig. 3A and data not shown).
However, these inactive siRNAs were able to compete with the
silencing of the active siRNAs. In these competition experiments,
inactive CD8 siRNAs were added into the cuvettes containing the
active CD8 siRNA, so that both could be electroporated into the
cells simultaneously. Varying concentrations were tested, and cells
were monitored for CD8 silencing at 36 h (Fig. 4). It was found
that when the total siRNA pool contained an inactive CD4 or
CD8 siRNA, then silencing mediated by an active siRNA was
markedly reduced (Fig. 4, A and B). These results mirror the ability
for active siRNAs to compete for other active siRNAs, a response
that we observed for attempting silencing of both CD4 and CD8
simultaneously (Fig. 4, C and D). The inability to silence both CD4
and CD8 simultaneously in the same cell might suggest that
siRNA-mediated RNAi is titratable, as has been described for silencing
using long dsRNAs in C. elegans (29).
To test whether the above siRNAs were also inactive in other
cell types, the CD4 and CD8 genes were expressed from CMVdriven
promoters in HeLa cells. The CD8 expression construct
contained two regions that corresponded to target sites for effective
and ineffective siRNAs in E10. In this assay, cationic lipid cotransfection
of the mouse CD4 and CD8 plasmid vectors was
performed with either the effective or noneffective CD8 siRNA.
When compared with the nonspecific siRNA control, CD8-specific
RNAi silencing was recapitulated in HeLa cells, and the ORFtargeted
siRNA was still ineffective at silencing (Fig. 5A). These
results suggested that the noneffective siRNA phenomenon is not
unique to the T cell line, but is likely a feature of either the siRNA
sequence, or more likely the mRNA. The concentration dependence
of the effective and ineffective siRNA was evaluated in the
HeLa cell assay. In this experiment, cationic lipid:siRNA complexes
were preformed and added to the cells as previously described
(13). The effective siRNA exhibited a concentration dependence;
however, the ineffective siRNAs remained inactive even
at the highest concentrations (Fig. 5B).
siRNA-mediated silencing in primary mouse T cells
To test whether primary cells are sensitive to siRNA-mediated
silencing, the CD4/CD8 siRNAs characterized above were used
to silence in primary mouse T cells taken from spleen. In theseognizes
OVA peptide in the context of MHC class II were isolated
from these mice are predominantly CD4; however, a small number
(15%) of CD8 cells exist in these mice. Efforts to transfect
and silence naive T cells were unsuccessful, but if the cells were
stimulated to divide by the cognate OVA peptide, CD4 and CD8
silencing could be accomplished similar to the E10 thymoma cell
line. Electroporation of CD4 siRNAs into activated primary T cells
resulted in an approximate 5-fold decrease in CD4 surface expression
compared with an unrelated siRNA control (Fig. 6). Costaining
for CD8 on the same cells demonstrated that the down-regulation
of CD4 was specific. The maximal degree of silencing was
reached at 48 h posttransfection. Later time points could not be
collected because of reduced cell viability after 72 h in culture.
Similarly, the subset of CD8-positive T cells electroporated with
CD8 siRNA exhibited a maximal 3.3-fold decrease in CD8 levels.
Furthermore, the degree of silencing in the sample population
with the alternate coreceptor (i.e., CD4 in a CD8 siRNA-treated
sample) verified that the RNAi response was specific (data not
shown). These results demonstrate that primary, mature T cells are
able to perform RNAi. The overall degree, kinetics, and specificity
of silencing of CD4 or CD8 in primary T cells was comparable
to that of the E10 cell line, further supporting the validity of using
this line to characterize T cell RNAi.Discussion
The CD4 and CD8 T cell surface glycoproteins are of central
importance to immune function and disease. We have quantitatively
tested the efficacy of a variety of siRNAs to suppress the
expression of these glycoproteins. Targeting the CD4 and CD8
markers was attractive since turnover of coreceptor message is
fairly rapid (12 h for CD8), and changes in surface expression
can be rapidly and easily assayed by flow cytometry. In this analysis
of two different genes, we observed that T cells and thymocytic
cell lines are amenable to siRNA-mediated silencing. These
studies revealed that siRNA-mediated RNAi is transient, lasting
approximately eight cell doublings. Not every siRNA was able to
induce silencing, and the RNAs which targeted the 3-UTR were
effective for both genes. Although small temporal RNAs (stRNAs)
mediated translational repression at the mRNA 3-UTR (for recent
reviews, see Refs. 30–34), Northern blot analysis of CD4 and
CD8 mRNA indicated posttranscriptional degradation of the
mRNA, consistent with a RNAi-type mechanism of silencing. Finally,
in primary T cells, the overall penetrance and kinetics of
CD4 and CD8 siRNA-mediated RNAi was found to be similar to
that observed in the E10 thymoma cell line.
In several experiments, and using electroporation, we found efficient
uptake and silencing of 90% of the cells. However, this
required the addition of a relatively high amount of siRNA (2.5
mol/1.5 107 cells); Northern blot analysis indicates that only a
fraction of the siRNAs (3 104 siRNAs/cell) become associated
with the cells (data not shown). Only a fraction of the siRNAs that
become associated with cells probably are functional in silencing
gene expression. At lower concentrations of siRNAs, a similar
fraction (70–95%) of cells exhibit a reduction in CD8 expression,
albeit at reduced efficiency. Using either electroporation for T cells
or Lipofectamine 2000 for HeLa cells, we found that 100% of the
cells that take up and express a cotransfected GFP marker also
perform RNAi. Based on this fact, it should be possible to design
gene function experiments which enrich the pool of silenced cells
by selecting for the activity of a transfected plasmid reporter.
Time course analysis of CD8 silencing in the E10 cell line
indicated that the silencing was transient in nature, lasting 3–4
days. As this cell line doubles rapidly, this value corresponds to
approximately eight cell doublings. Northern blots indicated that
silencing corresponded to a reduction in mRNA levels, commensurate
with the predicted model for RNAi. A translational repression
mechanism has been suggested for silencing mediated by
stRNAs via the 3 untranslated region of developmentally important
genes. Although the reduction in mRNA level approximated
that of CD8 expression, we cannot rule out the possibility of
additional translational repression mechanisms.
Only a limited number of the siRNA sequences tested could
induce RNAi. For the silencing of most genes, on average one of
two candidate siRNAs designed is active in contrast to the one of
four and one in five siRNAs tested in targeting CD4 and CD8 (6).
It is interesting to note that the siRNAs that were active in silencing
targeted the 3-UTR and stop codon. The restrictive utilization
of the 3-UTR siRNAs did not appear to be cell-type specific, as
active and inactive siRNAs gave similar results in HeLa cells. It is
unclear why targeting the mouse CD4 and CD8 mRNA 3-UTRs
were effective for performing siRNA-mediated RNAi, while other
sites were not. One possibility is that further testing of other
mRNA regions would result in productive silencing (35). Alternatively,
perhaps the 3-UTR of these genes is particularly accessible
for targeting. Silencing of developmentally timed genes in the endogenous
stRNA pathway is specific for the 3-UTR (25, 36). This
could be a common feature of developmentally timed genes, because
both CD4 and CD8 are also expressed in a developmentally
timed manner.
Attempting to silence both CD4 and CD8 simultaneously resulted
in lower levels of silencing of each gene. These results
supports a previously recognized observation that the RNAi response
is titratable (29). Surprisingly, several of the siRNAs that
were inactive competed for silencing when coelectroporated with
active siRNAs. While this manuscript was in preparation, another
group reported similar findings for the silencing of human coagulation
trigger factor (37). However, another group has reported
success in dual gene targeting of Lamin A/C and NuMA proteins
in HeLa cells (38). The data presented in this study indicate that
the inactive siRNAs are recognized by cellular processes but either
cannot be converted to an active structure for gene silencing or
cannot gain access to their complementary sequences on the
target mRNA.
This work presents the first evidence for silencing by siRNA in
primary somatic mammalian lymphocytes. In these studies, the
degree and kinetics of CD4 and CD8 silencing in the activated
primary cells was similar to that of the E10 cell line. In both the
primary cells and E10 cells the onset of maximal silencing appeared
around three to four cell doublings, which corresponded to
36–48 h posttransfection. In the E10 cells, 100% of the cells had
resumed normal CD8 expression by 96 h. Because the viability of
the primary cells began to diminish at around 60 h, it was difficult
to determine how long the RNAi response would last past 72 h. It
is interesting to note that the cells needed to be activated in order
for silencing to be accomplished. This could be due to the inability
to take up the siRNAs after electroporation, as primary T cells are
known to be difficult to transfect with nucleic acids. It is unknown
whether mammalian cells must be in a dividing, or “competent”,
state to perform RNAi. Future studies of siRNA-mediated RNAi in
primary cells are required to distinguish between these two possibilities.
Nevertheless, these findings provide a precedent upon
which future studies of T lymphocyte biology can be designed to
validate function by siRNA-mediated silencing.
这是一个关于小干扰技术在淋巴细胞的膜蛋白的上应用的文章，欢迎大家提出问题，共同讨论、共同提高！
相互交流、享受提高！！！
Small Interfering RNA-Mediated Gene Silencing in T
Lymphocytes1
Michael T. McManus, Brian B. Haines, Christopher P. Dillon, Charles E. Whitehurst,
Luk van Parijs, Jianzhu Chen, and Phillip A. Sharp2
Introduction of small interfering RNAs (siRNAs) into a cell can cause a specific interference of gene expression known as RNA
interference (RNAi). However, RNAi activity in lymphocytes and in normal primary mammalian cells has not been thoroughly
demonstrated. In this report, we show that siRNAs complementary to CD4 and CD8 specifically reduce surface expression of
these coreceptors and their respective mRNA in a thymoma cell line model. We show that RNAi activity is only caused by a subset
of siRNAs complementary to the mRNA target and that ineffective siRNAs can compete with effective siRNAs. Using primary
differentiated T lymphocytes, we provide the first evidence of siRNA-mediated RNAi gene silencing in normal nontransformed
somatic mammalian lymphocytes. The Journal of Immunology, 2002, 169: 5754–5760.
I ntroduction of dsRNA into an organism can cause specific
interference of gene expression (1). This phenomenon,
known as RNA interference (RNAi),3 results from a specific
targeting of mRNA for degradation by an incompletely characterized
cellular machinery present in plant, invertebrate, and mammalian
cells (2, 3). The proteins mediating RNAi are part of an
evolutionarily conserved cellular pathway that processes endogenous
cellular RNAs to silence developmentally important genes (4,
5). In RNAi, the protein Dicer, an RNase III enzyme, is probably
responsible for the processing of dsRNA into short interfering
RNA (siRNA). Functional screens conducted in plants and worms
have identified a number of other conserved genes participating in
the RNAi pathway. These genes include a number of different
helicases, a RNA-dependent RNA polymerase, an exonuclease,
dsRNA-binding proteins, and novel genes of unknown function
(for recent reviews, Refs. 6, 7, 8, 9, and 10).
Mammalian RNAi was first described in mouse embryos using
long dsRNA (11, 12). Then, following the analysis of the structure
of the intermediate in this process, small interfering RNAs
(siRNAs) were used to silence genes in mammalian tissue culture
(13, 14). Most of the RNAi pathway genes discovered in plant and
worm screens are also present in mouse and human sequence databases,
supporting evidence that a conserved RNAi pathway exists
in mammals. One of the more notable exceptions is the RNAdependent
RNA polymerase gene, which has been shown to beinvolved in the amplification of the dsRNA in Caenorhabditis elegans
(15, 16). This might imply that perpetuation of the RNAi
response in mammals differs from that of lower organisms.
Recent reports have demonstrated gene silencing by siRNA in
mammalian cells (17–22). However, despite these initial reports,
many uncertainties remain concerning the mechanism, physiologic
relevance, and ubiquity of RNAi in mammalian cells. Although
studies in tumor cell lines have demonstrated siRNA-mediated
RNAi, it remains a major question as to whether primary cells
from fresh tissues can undergo the RNAi response. Furthermore,
little is known about the efficiency and longevity of siRNA-mediated
RNAi gene suppression. In this report, we provide fundamental
insight into the siRNA-mediated RNAi mechanism using a thymoma-
derived cell line model to demonstrate for the first time the
occurrence of RNAi in primary T lymphocytes.
Materials and Methods
Cell culture
E10 is an immature double-positive thymocyte line derived from a TCR-
and p53 double-mutant mouse of a mixed 129/Sv C57BL/6 background
as described (23). These cells, which proliferated vigorously, were maintained
at a maximal concentration of 2 106 cells/ml and were propagated
in complete medium: DMEM supplemented with 10% heat-inactivated
FCS, 2 mM L-glutamine, 100 U/ml penicillin, 100 g/ml streptomycin, and
50 M 2-ME. Cell culture of primary lymphocytes: cells from the spleen
and lymph nodes of DO11.10 TCR-transgenic mice (a generous gift from
Dr. C. London, University of California, Davis, CA) were activated for 3
days with 1 g/ml OVA peptide (residues 323–339) in RPMI medium
containing 10% FBS.
Transfection
For electroporations, 2.5 mol dsRNA and/or 20 g of pEGFP-N3 plasmid
(Clontech Laboratories, Palo Alto, CA) were added to prechilled 0.4-cm
electrode gap cuvettes (Bio-Rad, Hercules, CA). E10 cells (1.5 107)
were resuspended to 3 107 cells/ml in cold serum-free RPMI, added to
the cuvettes, mixed, and pulsed once at 300 mV, 975 F with a Gene
Pulser electroporator II (Bio-Rad). Cells were plated into 6-well culture
plates containing 8 ml of complete medium and were incubated at 37°C in
a humidified 5% CO2 chamber. Cell viability immediately after electroporation
was typically around 60%. For cationic lipid transfections, 2 g of
plasmid DNA and 100 nmol siRNAs were used per 106 cells, and transfection
followed manufacturer’s recommended protocol. Transfection of
primary lymphocytes: activated DO11.10 T cells were electroporated as
above, except that the cells were resuspended to 6 107 cells/ml in cold
serum-free RPMI and the pulse voltage was 310 mV. After electroporation,
the cells were put into four wells of a 24-well plate, each containing 1 mlof RPMI supplemented with 1 ng/ml IL-2 (BioSource International, Camarillo,
CA). siRNA oligos (Dharmacon, Lafayette, CO) used were as follows
(sense strand is given): effective CD4 siRNA, CD4 no. 4, (sense)
gagccauaaucucaucugadgdg, (anti-sense) ucagaugagauuauggcucdtdt; effective
CD8 siRNA, CD8 no. 4, (sense) gcuacaacuacuacaugacdtdt, (antisense)
gucauguaguaguuguagcdtdt; ineffective siRNAs, CD8 no. 1, (sense) gaaaa
uggacgccgaacuudgdg, (anti-sense), aaguucggcguccauuuucdtdt; CD8 no. 2,
(sense) cgugggacgagaagcugaadtdt, (antisense) uucagcuucucgucccacgdtdt;
CD8 no. 3 (sense) aauuguguaaaauggcaccgcdcda, (antisense) ggcggugc
cauuuuacacaadtdt; CD4 no. 1, (sense) ggagaccaccaugugccgadgdc, (antisense)
ucggcacaugguggucuccdtdt; CD4 no. 2, (sense) ggcagagaaggauucu
uucdtdt, (anti-sense) gaaagaauccuucucugccdtdt; CD4 no. 3, (sense)
ccaccugcguccugucucadtdc, (antisense) gugguggacgcaggacagadgdt; CD4 no. 5
(sense) ccaccugcguccugucucadtdc, (antisense) ucagaugagauuauggcucdtdt.
Flow cytometry
E10 cells (1 106) were washed once in FACS buffer (PBS supplemented
with 2% FCS and 0.01% sodium azide), resuspended to 100 l, and
stained directly with PE-conjugated anti-CD4 (clone RM4-5) or allophycocyanin-
conjugated anti-CD8 mAbs, and in some experiments with PEor
allophycocyanin-conjugated anti-mouse Thy-1.2 (clone 53-2.1) mAb.
All mAbs were from BD PharMingen (San Diego, CA). The stained cells
were washed once, then resuspended in 200 l FACS buffer containing 200
ng/ml propidium iodide (PI). Unstained and singly stained controls were
included in every experiment. 3A9, a T cell hybridoma line that had been
infected with a MIGW green fluorescent protein (GFP) retrovirus was included
when GFP expression was analyzed. Cell data were collected on a
FACSCalibur flow cytometer (BD Biosciences, San Jose, CA) and fourcolor
analyses (GFP, PE, PI, and allophycocyanin) were done with
CellQuest software (BD Biosciences). All data were collected by analyses
performed on 1 104 PI-negative events (viable cells). For the primary T
cell studies, activated cells were analyzed as above, except that allophycocyanin-
conjugated anti-CD4 and PE-conjugated anti-CD8 were used,
and 5 104 PI-negative events were analyzed.
Northern blot analysis of mRNA
Cells were lysed in TRIzol reagent (Life Technologies, Grand Island, NY)
and total cellular RNA was purified according to manufacturer’s instructions.
RNA (10 g) was fractionated on a denaturing 1% formaldehyde/
agarose gel and transferred to a nitrocellulose membrane. Blots were hybridized
overnight with 32P-labeled CD4 (818 bp) or CD8 (596 bp) cDNA
fragments. After washes, blots were analyzed by a PhosphorImager (Molecular
Dynamics).
Results
siRNAs transiently induce silencing in murine thymocyte
cell lines
To study RNAi, siRNAs are typically delivered into cells by carrier-
mediated transfection reagents. We developed an experimental
system using a thymoma-derived cell line, E10 (23), wherein we
studied the use of siRNAs to silence either CD4 or CD8, using
the other marker as an internal specificity control. However, typical
of lymphocytes, E10 is insensitive to several different cationic
and noncationic transfection reagents and thus electroporation was
used to introduce siRNAs. Using this method, 20% of the cell
population expressed GFP from a transfected reporter vector.
When CD4 or CD8 siRNAs were electroporated into E10, a
marked reduction in surface CD4 or CD8 expression, respectively,
occurred 36 h later. Flow cytometry analysis showed that
most of the cells were transfected and expression levels were reduced
5-fold below wild-type expression levels (Fig. 1A). The
degree of reduction of CD8 was frequently more pronounced
than that of CD4 and, in both cases, a small population of cells
appeared to be either untransfected or not responsive to the siRNA
treatment. In repeated experiments, typically 70–95% of the cells
exhibited a 5-fold reduction in CD8 expression, although
sometimes a smaller fraction of cells down-regulated CD8 to a
greater degree (Fig. 1A).
Elbashir et al. (13) reported that RNAi-induced silencing could
be maintained for 2 wk in HeLa cells, although neither the extent
of silencing nor the number of cell divisions was reported. A timecourse assay was performed in CD8 siRNA-transfected E10
cells. GFP was included in these transfections to investigate the
relationship between the uptake and expression of plasmid DNA
and siRNAs. Because these experiments were transient transfections,
cell doubling results in a decrease in GFP fluorescence intensity
and number of GFP-positive cells (Fig. 1B, NS RNA).
When CD8 siRNAs were cotransfected with the GFP reporter
vector, CD8 expression, but not GFP expression, was markedly
reduced (see Fig. 1B, 24 h). Several cell populations were evident,
with the major CD8 silenced population displaying 5-fold reduced
CD8 expression. The majority of cells within this population
did not express GFP. However, cells that did express GFP
also silenced CD8. This corresponded to 20% of the total cells,
similar to the control GFP alone (Fig. 1B, NS siRNA, 24 h). This
indicates that all of the cells expressing GFP also received an adequate
level of siRNAs to silence CD8. In addition, a large fraction
of cells incorporated biologically active levels of siRNAs andyet did not express plasmid DNA. In this experiment, time points
were taken over a period of 6 days. At each time point, one-half of
the cells were removed from the dish and replaced with fresh medium.
The collected cells were stained for CD8 and analyzed by
flow cytometry (Fig. 1). A decrease in CD8 surface expression
was detectable at 12 h posttransfection, with maximal silencing at
36 h. By 96 h, nearly all of the cells expressed wild-type levels of
CD8. Thus, the RNAi effect in these T cells is a transient
phenomena.
In these experiments, there was a dramatic decrease in GFP
expression over time, which was likely a result of dilution of the
plasmid or potentially due to toxicity of high GFP expression.
Because 100% of the GFP-expressing cells exhibited CD8 silencing,
it was possible to monitor the “fate” of this subset of
silenced cells. The T cells that actively underwent CD8 silencing
continued to express GFP over the time course, to the same level
as the control population of cells that were not transfected with
siRNAs (compare nonspecific RNA to CD8 siRNA). At 96 h,
5% of the total cells were GFP-positive in cells treated with
nonspecific siRNAs and in CD8 siRNA-treated samples. These
few remaining GFP-positive cells exhibited normal levels of
CD8 expression. This suggests that the cells did not specifically
undergo apoptosis as a result of siRNA transfection and subsequent
CD8 silencing.
Specificity of siRNA-mediated silencing
Although the GFP transgene expression was not affected during
CD8 silencing, the expression of endogenous genes might have
been nonspecifically affected. To address this question, the expression
levels of CD4 and Thy1.2 T cell markers were examined in
cells actively undergoing CD8 silencing. Examination of these
markers revealed that there was no reduction of nontargeted gene
expression when compared with the control nontransfected cells
(Fig. 2A), even over extended times (not shown). Although unlikely
for this cell line, an additional analysis confirmed that the T
cells did not become activated, as they do not up-regulate CD69
(Fig. 2A). Together, these experiments confirm the specificity of
siRNA-mediated CD8 silencing.
Stability of targeted CD8 mRNA
Short temporal RNAs such as lin-4 and let-7 mediate silencing by
binding to the 3-untranslated region (UTR), thus suppressing
translation (24–26). This is in marked contrast to the posttranscriptional
mRNA degradation effected by siRNAs. To distinguish between
these two potential mechanisms for CD8 silencing, a time
course Northern blot analysis of CD8 mRNA was performed.
The process of silencing did not appreciably affect the growth rate,
as compared with control nonspecific siRNA transfections performed
in parallel (not shown). Flow cytometry analysis indicated
that the RNAi response in these cells lasted 3–4 days (8–10 cell
doublings), which corresponds to an 100-fold increase in cell
mass (Fig. 2B). Time course analysis was performed in four independent
experiments and expression of CD8 was typically suppressed
5-fold or greater.
At various time points, a fraction of the cells was used to isolate
total RNA for Northern blot analysis (Fig. 2C). The CD8 mRNA
was resolved into two bands, due to alternative splicing (27, 28).
Levels of CD8 mRNA decreased during the course of CD8
silencing. Densitometric analysis of the CD8 mRNA bands was
performed and normalized to the internal control CD4 band. At the
point of maximal silencing, mRNA levels decrease only 2.5-fold.
This value is not commensurate with the 5-fold decrease in protein
expression determined by the flow cytometric analysis. However,
this RNA was prepared from total cells in which 30% of the
cells did not exhibit any silencing. When corrected for this reduction,
CD8 mRNA was nearly proportionate to levels in reduction
of CD8 protein. These Northern blots were performed multiple
times with similar results. Thus, although it is clear that CD8
mRNA decreases, we cannot rule out additional silencing phenomena
such as cotranslational repression.
Regional sensitivity of an mRNA to silencing by a siRNA
A major outstanding question is whether any region of a mRNA
can serve as an effective target for siRNA-directed silencing. Several
different siRNAs that targeted different regions of the CD8阿尔法mRNA were tested. Of the first two CD8 siRNAs that were transfected,
only one was active. To more quantitatively examine this
difference, cells were transfected with varying amounts of siRNAs
and CD8 expression was measured by flow cytometry. Cells undergoing
silencing were quantified and compared with control
nonspecific siRNA treatment (Fig. 3A). For the effective CD8
siRNA, picomolar amounts were sufficient to induce some silencing
and higher amounts produced a graded response. For the noneffective
CD8 siRNA, even at the highest concentration tested,
there was no activity.
As these studies progressed, we observed that the majority of the
synthetic CD4 and CD8 siRNAs were noneffective at silencing.
For CD8, four different siRNAs were synthesized and tested in
the flow cytometry assay: one overlapped the start codon, one
which targeted the open reading frame (ORF), one which overlapped
the stop codon, and one which targeted the 3-UTR 15 nt
after the stop codon. Only the siRNA which targeted the 3-UTR
15 nt after the stop codon effectively silenced CD8 expression.
For CD4, five siRNAs were synthesized which targeted corresponding
regions to those for the CD8 mRNA (Fig. 3B). In this
case, only the siRNA that targeted the stop codon was effective at
reducing CD4 expression levels. An examination of the nucleotide
sequences did not reveal any obvious differences between the effective
and ineffective siRNAs.
For each of the above siRNAs, the silencing assay was performed
at different siRNA concentrations. None of the inactive
siRNAs generated detectable silencing at five times the highest
concentration of the active siRNAs (Fig. 3A and data not shown).
However, these inactive siRNAs were able to compete with the
silencing of the active siRNAs. In these competition experiments,
inactive CD8 siRNAs were added into the cuvettes containing the
active CD8 siRNA, so that both could be electroporated into the
cells simultaneously. Varying concentrations were tested, and cells
were monitored for CD8 silencing at 36 h (Fig. 4). It was found
that when the total siRNA pool contained an inactive CD4 or
CD8 siRNA, then silencing mediated by an active siRNA was
markedly reduced (Fig. 4, A and B). These results mirror the ability
for active siRNAs to compete for other active siRNAs, a response
that we observed for attempting silencing of both CD4 and CD8
simultaneously (Fig. 4, C and D). The inability to silence both CD4
and CD8 simultaneously in the same cell might suggest that
siRNA-mediated RNAi is titratable, as has been described for silencing
using long dsRNAs in C. elegans (29).
To test whether the above siRNAs were also inactive in other
cell types, the CD4 and CD8 genes were expressed from CMVdriven
promoters in HeLa cells. The CD8 expression construct
contained two regions that corresponded to target sites for effective
and ineffective siRNAs in E10. In this assay, cationic lipid cotransfection
of the mouse CD4 and CD8 plasmid vectors was
performed with either the effective or noneffective CD8 siRNA.
When compared with the nonspecific siRNA control, CD8-specific
RNAi silencing was recapitulated in HeLa cells, and the ORFtargeted
siRNA was still ineffective at silencing (Fig. 5A). These
results suggested that the noneffective siRNA phenomenon is not
unique to the T cell line, but is likely a feature of either the siRNA
sequence, or more likely the mRNA. The concentration dependence
of the effective and ineffective siRNA was evaluated in the
HeLa cell assay. In this experiment, cationic lipid:siRNA complexes
were preformed and added to the cells as previously described
(13). The effective siRNA exhibited a concentration dependence;
however, the ineffective siRNAs remained inactive even
at the highest concentrations (Fig. 5B).
siRNA-mediated silencing in primary mouse T cells
To test whether primary cells are sensitive to siRNA-mediated
silencing, the CD4/CD8 siRNAs characterized above were used
to silence in primary mouse T cells taken from spleen. In theseognizes
OVA peptide in the context of MHC class II were isolated
from these mice are predominantly CD4; however, a small number
(15%) of CD8 cells exist in these mice. Efforts to transfect
and silence naive T cells were unsuccessful, but if the cells were
stimulated to divide by the cognate OVA peptide, CD4 and CD8
silencing could be accomplished similar to the E10 thymoma cell
line. Electroporation of CD4 siRNAs into activated primary T cells
resulted in an approximate 5-fold decrease in CD4 surface expression
compared with an unrelated siRNA control (Fig. 6). Costaining
for CD8 on the same cells demonstrated that the down-regulation
of CD4 was specific. The maximal degree of silencing was
reached at 48 h posttransfection. Later time points could not be
collected because of reduced cell viability after 72 h in culture.
Similarly, the subset of CD8-positive T cells electroporated with
CD8 siRNA exhibited a maximal 3.3-fold decrease in CD8 levels.
Furthermore, the degree of silencing in the sample population
with the alternate coreceptor (i.e., CD4 in a CD8 siRNA-treated
sample) verified that the RNAi response was specific (data not
shown). These results demonstrate that primary, mature T cells are
able to perform RNAi. The overall degree, kinetics, and specificity
of silencing of CD4 or CD8 in primary T cells was comparable
to that of the E10 cell line, further supporting the validity of using
this line to characterize T cell RNAi.Discussion
The CD4 and CD8 T cell surface glycoproteins are of central
importance to immune function and disease. We have quantitatively
tested the efficacy of a variety of siRNAs to suppress the
expression of these glycoproteins. Targeting the CD4 and CD8
markers was attractive since turnover of coreceptor message is
fairly rapid (12 h for CD8), and changes in surface expression
can be rapidly and easily assayed by flow cytometry. In this analysis
of two different genes, we observed that T cells and thymocytic
cell lines are amenable to siRNA-mediated silencing. These
studies revealed that siRNA-mediated RNAi is transient, lasting
approximately eight cell doublings. Not every siRNA was able to
induce silencing, and the RNAs which targeted the 3-UTR were
effective for both genes. Although small temporal RNAs (stRNAs)
mediated translational repression at the mRNA 3-UTR (for recent
reviews, see Refs. 30–34), Northern blot analysis of CD4 and
CD8 mRNA indicated posttranscriptional degradation of the
mRNA, consistent with a RNAi-type mechanism of silencing. Finally,
in primary T cells, the overall penetrance and kinetics of
CD4 and CD8 siRNA-mediated RNAi was found to be similar to
that observed in the E10 thymoma cell line.
In several experiments, and using electroporation, we found efficient
uptake and silencing of 90% of the cells. However, this
required the addition of a relatively high amount of siRNA (2.5
mol/1.5 107 cells); Northern blot analysis indicates that only a
fraction of the siRNAs (3 104 siRNAs/cell) become associated
with the cells (data not shown). Only a fraction of the siRNAs that
become associated with cells probably are functional in silencing
gene expression. At lower concentrations of siRNAs, a similar
fraction (70–95%) of cells exhibit a reduction in CD8 expression,
albeit at reduced efficiency. Using either electroporation for T cells
or Lipofectamine 2000 for HeLa cells, we found that 100% of the
cells that take up and express a cotransfected GFP marker also
perform RNAi. Based on this fact, it should be possible to design
gene function experiments which enrich the pool of silenced cells
by selecting for the activity of a transfected plasmid reporter.
Time course analysis of CD8 silencing in the E10 cell line
indicated that the silencing was transient in nature, lasting 3–4
days. As this cell line doubles rapidly, this value corresponds to
approximately eight cell doublings. Northern blots indicated that
silencing corresponded to a reduction in mRNA levels, commensurate
with the predicted model for RNAi. A translational repression
mechanism has been suggested for silencing mediated by
stRNAs via the 3 untranslated region of developmentally important
genes. Although the reduction in mRNA level approximated
that of CD8 expression, we cannot rule out the possibility of
additional translational repression mechanisms.
Only a limited number of the siRNA sequences tested could
induce RNAi. For the silencing of most genes, on average one of
two candidate siRNAs designed is active in contrast to the one of
four and one in five siRNAs tested in targeting CD4 and CD8 (6).
It is interesting to note that the siRNAs that were active in silencing
targeted the 3-UTR and stop codon. The restrictive utilization
of the 3-UTR siRNAs did not appear to be cell-type specific, as
active and inactive siRNAs gave similar results in HeLa cells. It is
unclear why targeting the mouse CD4 and CD8 mRNA 3-UTRs
were effective for performing siRNA-mediated RNAi, while other
sites were not. One possibility is that further testing of other
mRNA regions would result in productive silencing (35). Alternatively,
perhaps the 3-UTR of these genes is particularly accessible
for targeting. Silencing of developmentally timed genes in the endogenous
stRNA pathway is specific for the 3-UTR (25, 36). This
could be a common feature of developmentally timed genes, because
both CD4 and CD8 are also expressed in a developmentally
timed manner.
Attempting to silence both CD4 and CD8 simultaneously resulted
in lower levels of silencing of each gene. These results
supports a previously recognized observation that the RNAi response
is titratable (29). Surprisingly, several of the siRNAs that
were inactive competed for silencing when coelectroporated with
active siRNAs. While this manuscript was in preparation, another
group reported similar findings for the silencing of human coagulation
trigger factor (37). However, another group has reported
success in dual gene targeting of Lamin A/C and NuMA proteins
in HeLa cells (38). The data presented in this study indicate that
the inactive siRNAs are recognized by cellular processes but either
cannot be converted to an active structure for gene silencing or
cannot gain access to their complementary sequences on the
target mRNA.
This work presents the first evidence for silencing by siRNA in
primary somatic mammalian lymphocytes. In these studies, the
degree and kinetics of CD4 and CD8 silencing in the activated
primary cells was similar to that of the E10 cell line. In both the
primary cells and E10 cells the onset of maximal silencing appeared
around three to four cell doublings, which corresponded to
36–48 h posttransfection. In the E10 cells, 100% of the cells had
resumed normal CD8 expression by 96 h. Because the viability of
the primary cells began to diminish at around 60 h, it was difficult
to determine how long the RNAi response would last past 72 h. It
is interesting to note that the cells needed to be activated in order
for silencing to be accomplished. This could be due to the inability
to take up the siRNAs after electroporation, as primary T cells are
known to be difficult to transfect with nucleic acids. It is unknown
whether mammalian cells must be in a dividing, or “competent”,
state to perform RNAi. Future studies of siRNA-mediated RNAi in
primary cells are required to distinguish between these two possibilities.
Nevertheless, these findings provide a precedent upon
which future studies of T lymphocyte biology can be designed to
validate function by siRNA-mediated silencing.
这是一个关于小干扰技术在淋巴细胞的膜蛋白的上应用的文章，欢迎大家提出问题，共同讨论、共同提高！
