RNA-Seq analysis of early transcriptional responses to activation in the leukaemic Jurkat E6.1 T cell line

Background: The leukaemia-derived Jurkat E6.1 cell line has been used as a model T cell in the study of many aspects of T cell biology, most notably activation in response to T cell receptor (TCR) engagement. Methods: We present whole-transcriptome RNA-Sequencing data for Jurkat E6.1 cells in the resting state and two hours post-activation via TCR and CD28. We compare early transcriptional responses in the presence and absence of the chemokines CXCL12 and CCL19, and perform a basic comparison between observed transcriptional responses in Jurkat E6.1 cells and those in primary human T cells using publicly deposited data. Results: Jurkat E6.1 cells have many of the hallmarks of standard T cell transcriptional responses to activation, but lack most of the depth of responses in primary cells. Conclusions: These data indicate that Jurkat E6.1 cells hence represent only a highly simplified model of early T cell transcriptional responses.


Introduction
Adaptive immunity is centred on the clonal selection and activation of lymphocytes, most importantly T cells, which provide stimulation and regulation to other cells of the immune system as well as directly killing infected cells. T cells become activated in response to binding of their clonally specific T cell receptors (TCRs) to cognate peptide-major histocompatibility complexes (pMHCs) on the surface of interacting antigenpresenting cells. This leads to recruitment and phosphorylation of a series of kinases and adaptor proteins to the TCR, intracellular Ca 2+ mobilisation, and consequent downstream changes in gene expression -largely due to activation of nuclear factor of activated T cells (NFAT) -which mark the transition to a fully activated state. Much of our understanding of the early events of T cell activation has been facilitated by the use of a small number of T cell lines, due largely to their relative ease of handling, manipulation, and transfection. Arguably the most widely used such case is the Jurkat cell line, which was originally isolated from the peripheral leukaemic T cells of a 14-year-old boy in the late 1970s (Abraham & Weiss, 2004). The E6.1 clone was subsequently derived from this due to its high capacity for interleukin-2 (IL2) secretion, and rapidly became the standard line for many prominent T cell biologists at the time.
The contribution of the Jurkat E6.1 line to T cell research is unquestionable and it is still widely used despite greatly improved technologies for ex vivo manipulation of primary T cells. Nonetheless, it is well described that some significant differences exist between the activatory processes and outcomes in primary and Jurkat T cells; an unavoidable consequence of the abnormal origins of the line. For example, Jurkats exhibit elevated phosphorylation of several core signalling proteins, such as phospholipase C γ1 (PLCγ1) and extracellular signalrelated kinases (ERKs) 1 and 2 ( Bartelt et al., 2009). This is linked in part to constitutive phosphatidylinositol 3-kinase (PI3K) activity due to a defect in phosphatase and tensin homolog (PTEN) expression (Shan et al., 2000), which also leads to hypersensitivity to TCR engagement compared to primary T cells. Relative to primary cells, Jurkats also exhibit differences in their cytoskeletal dynamics and organisation (Colin-York et al., 2019), which may influence their response to activation due to effects on the formation of the immunological synapse (Dustin, 2014). Hence, although Jurkats have retained many of the core aspects of normal T cell signalling, there are many features of their biology that deviate substantially from primary cells.
One complex aspect of T cell activation is how signals deriving from the TCR become integrated with those from other modulatory receptors. An interesting example of this is the contribution of chemokines to T cell activation, given that chemokines are more usually associated with cell migration ( Here, we present whole transcriptome RNA-Sequencing (RNA-Seq) data for Jurkat E6.1 cells under resting conditions and two hours post-activation in the presence and absence of the chemokines CXCL12 and CCL19, which ligate the established costimulatory receptors CXCR4 and CCR7, respectively. In order to probe the differences in early transcriptional responses between Jurkats and primary T cells, we compare our data to publicly deposited data derived from equivalently activated primary T cells.
Stimulation was performed using anti-human CD3/CD28 Dynabeads (Gibco) in complete growth medium. Cells were centrifuged at 300 × g for 5 min then resuspended in warm growth medium at 2.5 × 10 6 /ml. Next, 2 ml/well of cell suspension was added to 3 wells of a 6-well plate and equilibrated at 37°C, 5% CO 2 for 1 h. Following this, 2 × 10 7 anti-CD3/CD28 beads were washed with growth medium and resuspended in 200 μl growth medium. Then, 100 μl of beads were then added to 2 wells and gently mixed, the other well was left as the resting condition. After 5 min, CXCL12 (PeproTech) and CCL19 (PeproTech) were added to one well to a final concentration of 100 ng/ml each. Cells were returned to the incubator for 2 h (for RNA-Seq experiments) or 2, 4, 6, or 24 h (for flow cytometry experiments), before removal and downstream use.

Amendments from Version 1
This version of the manuscript incorporates a number of technical clarifications and further points of discussion, as suggested by the reviewers. We have swapped the numbering of Figure 5 and Figure � � as we realised that the order they were mentioned in the text was incorrect. Additionally, a further figure (Figure 7) has been added. It also includes a comparison between resting Jurkat data and public single cell RNA-Seq data of different primary CD4+ subsets, including both effector and central memory cells in order to assess which primary T cell population exhibits greatest transcriptional similarity with Jurkat cells.
Any further responses from the reviewers can be found at the end of the article REVISED RNA isolation and RNA-Seq For each condition, total RNA was isolated from 5 × 10 6 cells, which were centrifuged at 300 × g for 5 min then resuspended in 2 ml TRIzol® reagent (ThermoFisher), incubated at room temperature for 10 min before adding 0.4 ml chloroform. The cell suspension was mixed thoroughly then centrifuged at 5,000 × g, 4°C for 30 min. The top aqueous layer (~800 μl) was carefully removed and replaced with an equivalent volume of isopropanol. Samples were then centrifuged at 17,000 × g, 4°C for 30 min to pellet the total RNA. Isopropanol was removed and the pellet washed with 1 ml EtOH, then centrifuged again at 17,000 × g, 4°C for 5 min. EtOH was removed and the pellet was air-dried for 15 min then resuspended in 200 μl RNase-free H 2 O. To remove contaminant DNA, 22 μl 10× TURBO DNase buffer (ThermoFisher) and 2 μl (4U) TURBO DNase (ThermoFisher) were added the RNA suspension then incubated at 37°C for 30 min. The described TRIzol®chloroform extraction protocol was then repeated to remove the DNase, and the final RNA sample resuspended in 200 μl RNase-free H 2 O. RNA was isolated twice for each condition in two independent experiments. ds-cDNA libraries for each sample were prepared and sequenced at the Wellcome Trust Centre for Human Genetics, Oxford, using the HiSeq ® 4000 Sequencing System (Illumina). Comparison of Jurkat gene expression with that of different primary CD4+ subsets was performed by comparison to publicly deposited single cell RNA-Seq data. Primary human CD4+ gene expression was obtained from a published single cell RNA-Seq and CITE-Seq multimodal reference atlas of the circulating immune system (Hao et al., 2020). The H5 Seurat data file was downloaded from https://atlas.fredhutch.org/nygc/multimodal-pbmc/ and loaded into R Studio (version 4.0.1) using Seurat package version 4.0.1. The counts matrix and associated metadata were extracted and converted to a Single Cell Experiment using SingleCellExperiment package version 1.10.1 (Amezquita et al., 2020). Pseudobulk counts were generated by aggregating counts for each donor (8 in total) and CD4+ T cell subtype: CD4 cytotoxic T lymphocytes (CTL), CD4 naive, CD4 proliferating, CD4 T central memory (TCM), CD4 T effector memory (TEM) and regulatory T cell (Treg). Both pseudobulk and Jurkat counts matrices were pre-processed and merged resulting in 11,705 genes. A DESeq2 object was created using the final counts matrix and associated metadata using DESeq2 version 1.28.1. Sample similarity/dissimilarity was visualised using principal component analysis (PCA) and hierarchal clustering (using pheatmap version 1.0.12).

RNA-Seq data analysis
Primary CD4 + T cell isolation and culture Primary human CD4 + T cells were isolated using the RosetteSep Human CD4 + T Cell Enrichment Cocktail (Stem-Cell Technologies) as per the manufacturer's instructions from leukocyte cones provided by UK National Health Service Blood and Transplant. Isolated cells were cultured in RPMI-1640 supplemented with 10% FCS (Gibco), 4 mM L-glutamine (Gibco), 10 mM HEPES (Gibco), 1% non-essential amino acid solution (Gibco), and 1% penicillin-streptomycin solution (Gibco) at 37°C, 5% CO 2 for between 24 and 72 h before stimulating in the same manner as for Jurkat cells.

Statistical analysis
The prcomp function was used for the principal component analysis and plotPCA was used for principal component analysis and visualisation; both are part of DESeq2 package. Intensity values derived from flow cytometry data were analysed and visualised using GraphPad Prism, version 8.2.1.

Results
Jurkat responses broadly correspond to expected effects of T cell activation and are unaffected by the presence of chemokines RNA-Seq was performed on total mRNA collected from Jurkat E6.1 cells under both resting and activated conditions (total mapped counted given in Extended data -Dataset S1 (Felce, 2020b)). Activation was performed using anti-CD3, anti-CD28 beads in the absence of exogenous IL2, and with either 0 or 100 ng/ml soluble CXCL12 and CCL19. Cells were activated for 2 h before lysis and mRNA extraction. This time point was used in order to interrogate only early transcriptional responses that occur during the expected lifetime of the T cell immunological synapse (up to the order of several hours) and hence may influence events occurring within the contact. Sequencing yielded between 4x10 7 and 4.9x10 7 individual reads per sample, and were performed as biological replicates with duplicate mRNA samples isolated in independent experiments. Following mapping to the human GRCh37.EBVB95-8wt.ERCC reference genome, differential gene expression was determined using DESeq2. Multimapped reads were not included in downstream analysis.
Firstly, gene expression was compared between resting Jurkats and those activated in the absence of chemokines. As expected, expression levels of many genes with a wide range of mean read counts were substantially altered in response to cellular activation ( Figure 1A; full differential gene list given in Extended data -Dataset S2 (Felce, 2020b)). The majority of genes with significantly (p.adj<0.01) altered expression upon activation underwent upregulation, including numerous wellcharacterised markers of T cell activation, e.g. CD69, IL2, IL2RA, and IFNG ( Figure 1B). GO enrichment analysis revealed that GO terms associated with significantly differentially expressed genes were largely connected to T cell activation, differentiation, and/or adhesion ( Figure 1C; Extended data -Dataset S3 (Felce, 2020b)), as expected. The same comparison performed between resting Jurkats and those activated in the presence of soluble chemokines to receptors CXCR4 (CXCL12) and CCR7 (CCL19) yielded comparable results, with no apparent effect of chemokines on differential gene expression in response to activation ( Figure 1D, E). Accordingly, when sample-to-sample and principal component variances are compared all samples within activated conditions cluster together due to high similarity, and are clearly distinct from samples in resting state (Figure 2A, B). Direct differential gene expression analysis between the activated samples in the presence and absence of chemokine yielded no significantly different genes ( Figure 2C, D).

The majority of early transcriptional responses observed in primary T cells are not replicated in Jurkats
Whilst the observed early transcriptional responses in Jurkats unquestionably correlate with known processes of T cell activation, we were interested in whether they deviated substantially from the situation in primary cells. We therefore took the opportunity to compare our dataset with publicly deposited RNA-Seq data of primary T cell activation. At present, no data are publicly available for primary human CD4 + T cells activated in an analogous manner for 2 h; however, a comparable analysis of activated primary CCR6 + CD4 + memory T cells has been previously published (Zhao et al., 2014). In this case the activatory conditions included Th17-polarising cytokines and antibodies, which were not present during the stimulation of Jurkats and hence may skew transcriptional responses accordingly. Nonetheless, this dataset allowed for a crude comparison between Jurkat and primary cells.
We examined unstimulated and 2 h stimulated raw read counts derived from the publicly deposited dataset (NCBI SRA: SRP026389) using the same analysis as for the Jurkat-derived data, then compared log2 fold change for all genes across both cell types and the fraction of significantly up-or downregulated genes. A far greater proportion of genes expressed in primary memory CD4 + cells (57%) had significantly altered expression upon stimulation compared to those in Jurkats (7%; Figure 3). Of those genes differentially expressed in Jurkats, 51% of upregulated and 33% of downregulated genes were similarly regulated in primary cells; however, the vast majority (86% and 98% of up-and downregulated genes, respectively) of genes with altered expression in primary cells exhibited no substantial change in Jurkats (Figure 3, C). As would be expected, genes that underwent upregulation in both cell types were strongly associated with GO terms linked to T cell activation, cell-cell adhesion, cytokine regulation, and other related biological processes ( Figure 3, Table 1; Extended data -Dataset S4 (Felce, 2020b)). Genes upregulated only in primary cells also strongly correlated with many such processes ( Figure 3, Table 1; Extended data -Dataset S4 (Felce, 2020b)), suggesting that Jurkats lack either expression or early activation-induced regulation of many genes associated with normal T cell responses. Examples of such genes span a wide range of functions, including inhibitory surface receptors (e.g. BTLA, CTLA4, TIGIT, LAG3), regulatory ligands (e.g. CD40, CD80, CD86, PDL1), adhesion molecules (e.g. ICAM1, CD2, CD58), cytokines (e.g. TGF1B, IL4, IL6, IL10), chemokines (e.g. CCL19, CXCL1, CXCL2, CXCL3), G proteincoupled receptors (e.g. CNR1, PTGER1, GPR183, GPR18), and transcription factors (e.g. FOXP1, FOXP3, IRF8, CBFB). The relatively small number of genes upregulated in Jurkats but not primary cells exhibited weak association to six biological processes related to the regulation of protein phosphorylation or synaptic transmission ( Figure 3H, Table 1; Extended data -Dataset S4 (Felce, 2020b)). Shared and non-shared downregulated genes across both cell types did not generally correlate significantly with known biological processes, except in the case of primary-only downregulated genes, which included a number of Toll-like receptors and associated proteins (Extended data -Dataset S4 (Felce, 2020b)).
Some of the differences between the transcriptional responses of the two cell types is likely due to the detection of genes restricted to the memory CD4 + lineage but not to effector CD4 + cells, or may result from Th17 polarisation. Nonetheless, the prevalence of so many known activation-associated genes in the primary-only subset indicates a genuine disparity in transcriptional responses between Jurkat and primary T cells. To test this, Jurkat and total primary CD4 + T cells were stimulated in the same way as for the RNA-Seq experiments for 2, 4, 6, and 24 h, stained with conjugated antibodies against CD69 and CD25 (which exhibited mRNA upregulation in both cell types), and BTLA, CTLA4, and CD86 (which exhibited mRNA upregulation only in the primary cell data) then assessed by flow cytometry. All five proteins exhibited detectable upregulation following 24 h stimulation in primary cells (with increases in CTLA4, CD86, and CD69 detectable following 2 h stimulation), whereas Jurkats only demonstrated increased expression of CD69 and CD25, in line with the RNA-Seq data (Figure 4). This indicates that although there is probably some skewing of the primary cell RNA-Seq data towards the memory and Th17 phenotypes, many of the observed differences with Jurkat transcriptional responses are likely to also hold true in conventional CD4 + T cells under non-polarising conditions. Although no comparable public datasets exist for non-memory primary CD4 + T cells activated for 2 h, data are available for total CD4 + T cells following 24 h activation: GEO: GSE122735 (Lucic et al., 2019). This allowed comparison with the same RNA-seq data derived from memory CD4 + cells used earlier but only at a later stage of activation. At this time point, there was strong correlation between log2 fold change relative to unstimulated cells across both primary cell types ( Figure 3I). A much greater proportion of up-and downregulated genes were shared between the two cell types than between memory T cells and Jurkats ( Figure 3J). Genes upregulated only in the total but not memory cells were associated largely with cell division, whilst those upregulated only in memory cells related primarily to Th17 differentiation due to the polarising conditions present during stimulation (Extended data -Dataset S5 (Felce, 2020b)). Memory cells, but not total CD4 + T cells, exhibited downregulation of several genes associated with T cell activation and cell-cell interactions (Extended data -Dataset S5 (Felce, 2020b)). In all, this indicates that although there are differences between the memory and total CD4 + T cell responses, the majority of observed transcriptional changes correlate across the two cell types. This lends confidence that the comparison described above for Jurkats and memory CD4 + cells at 2 h stimulation is not inappropriate. This is supported by the observation that there was comparably poor correlation between the 2 h activated Jurkat dataset and both the total and memory primary CD4 + T cells activated for 24 h, whereas there was much greater correlation between the 2 h and 24 h activated memory ( Figure 5).

Differences exist between baseline transcriptome in
Jurkats and primary CD4 + cells In order to confirm the Jurkat line from which we derived these data was not abnormal relative to other Jurkat E6.1 cells  H). Only 6 significant terms were identified for Jurkat-only genes. Terms associated with T cell activation are enriched in both shared and memory CD4 + -only sets, but not Jurkat-only sets. (I) Log2 fold change for all genes across resting and 24 h activated conditions in primary memory CD4 + vs. total CD4 + T cells. (J) Venn diagrams of total upregulated (log2 fold change > 1) and downregulated (log2 fold change < -1) genes shared and non-shared between primary memory CD4 + and total CD4 + T cells.  Figure 6B). Baseline transcription in the CCR6 + memory T cell dataset used in the earlier comparisons (SRP026389) was almost entirely (98%) shared with at least one total CD4 + T cell dataset ( Figure 6B).
We took this opportunity to broadly compare the baseline transcriptomes of Jurkat and primary CD4 + T cells. In total, 15,061 genes were shared between our and both public Jurkat datasets, while 16,472 genes were shared between both total primary CD4 + T cell datasets. The majority of these genes were shared across both cell types, however both also exhibited substantial fractions of genes that were cell type-specific (14% in Jurkats, 21% in primary CD4 + cells; Figure 6C). Many genes expressed in primary CD4 + T cells but not Jurkats associated closely with several GO terms connected to T cell responses; predominantly innate immune activation, cytokine secretion, and leukocyte adhesion/migration (Table 2, Figure 6D; Extended data -Dataset S6 (Felce, 2020b)). By contrast, Jurkat-restricted genes did not generally associate with immune function but instead with a range of developmental and/or neurological biological processes (Table 2, Figure 6E; Extended data -Dataset S6 (Felce, 2020b)). This again indicates that Jurkats do not fully reproduce the transcriptional state of primary T cells, and hence there may be deviations in the activationinduced transcriptional changes between the two cell types.
To further assess the deviation of the Jurkat transcriptome from those of primary cells, we compared our resting Jurkat dataset with publicly deposited single cell RNA-Seq data for six different primary CD4 + T cell populations: naïve, proliferating, central memory, effector memory, regulatory, and cytotoxic. This revealed that the transcriptome diversity between different primary CD4+ subtypes was markedly less than that between Jurkats and any single primary population (Figure 7). Proliferating CD4+ T cells showed the smallest difference from Jurkats, most likely due to the expression of genes involved in cell cycle progression. These data provide confidence that the memory CD4+ subset used in the earlier comparison with activated Jurkats is most likely comparably transcriptionally appropriate as other primary CD4+ T cells.

Discussion
The Jurkat E6.1 cell line has been extensively used as a model of T cell biology, particularly in the study of TCR signalling.
Here we have presented RNA-Seq data revealing the early transcriptional effects of Jurkat activation through engagement of the TCR and CD28. We have observed that transcriptional modulation following 2 h of stimulation correlates closely with many of the known responses to activation in T cells, and that this is unaffected by the presence of the chemokines CXCL12 and CCL19. Nonetheless, when compared to publicly available data for primary T cells stimulated in the same manner it is evident that these changes represent only a small fraction of the early responses occurring in primary cells. Many genes lacking modulation in Jurkats are closely associated with known pathways of T cell activation, and a subset of these was validated at the protein level using flow cytometry. This is underlined further by substantial differences in the baseline, resting transcriptomes of Jurkat and primary T cells.  identical in both cases due to the presence of Th17-polarising chemokines with the primary T cells, explaining the upregulation of several Th17-associated genes (e.g. IL17F) in the primary sample. Nonetheless, the core activatory signals were the same in both cases, and so it seems reasonable to expect that many central transcriptional responses should be conserved between both cell types. The lack of evident transcriptional regulation in Jurkats for a majority of genes altered in primary cells indicates a fundamental difference between the two cell types, in which the Jurkat transcriptome is much less responsive to early activatory signals.
It must also be considered that the primary memory CD4 + T cell data were derived from a single donor, which we are assuming is representative of normal primary T cell behaviour. There is some reassurance that this is the case from the 24 h stimulation condition, which exhibits good correlation with data derived from multiple donors of total CD4 + T cells ( Figure 3I); however, a donor-specific effect cannot be ruled out.
An alternative explanation for these observations is that transcriptional modulation in Jurkats is slower than in primary cells, and hence many apparently primary-specific changes in gene expression may be replicated in Jurkats at later times post-stimulation. There does not appear to be a general trend among typical T cell activation markers to exhibit greater log2 fold change in the primary cell data (e.g. CD69 3.71 in Jurkats vs. 7.54 in primary cells; IL2 6.71 vs. 9.38; IL2RA 5.44 vs. 2.74; IFNG 7.74 vs. 7.39), and for the proteins assessed directly by flow cytometry this also does not appear to be the case (Figure 4). Nonetheless, this cannot be fully excluded as a general effect.
In summary, these data indicate that, at least at a transcriptional level, Jurkat E6.1 cell responses are far more minimal than those in primary T cells, and hence Jurkats represent a highly simplified model of T cell transcriptional modulation. We fully acknowledge the limitations of the comparisons made, insofar as the differences existing between cell types and activation conditions; however, the extent of the observed transcriptional Venn diagram of total genes expressed in two publicly deposited datasets for resting total primary CD4 + T cells, and one for resting memory primary CD4 + T cells. (C) Venn diagram of total genes shared among all three datasets for Jurkats, and both datasets for total primary CD4 + T cells. (D and E) Enrichment map plots of Gene Ontology terms significantly enriched among genes expressed in resting Jurkat or primary CD4 + T cells, but not both.

Table 2. Gene Ontology (GO) terms associated with unique baseline transcriptomes of resting Jurkat and primary T cells.
Top 20 GO terms with greatest enrichment in genes shared between multiple datasets for Jurkats or primary CD4 + T cells but not between the two cell types. Redundant terms have been removed for clarity. A full list of significant terms is available in the Extended Data (Dataset S6). differences and the direct validation of several key examples lends confidence to the central conclusions reported here. We provide the present dataset for other researchers to use in more direct comparisons to test the robustness of our conclusions. Moreover, given that Jurkats and other leukaemic cell lines are frequently used as models of acute lymphoblastic leukaemia (ALL), these observations underline the divergence of these models from normal T cells. Since such cell lines have highly varied origins and mutational profiles, it is most appropriate to use several ALL models rather than possibly over-interpreting experiments from a single cell line. This project contains the following underlying flow cytometry data:
• E3 -Flow cytometry data form donor 3. This project contains the following extended data:

Extended data
• Dataset S1 -Raw count values for all samples.
• Dataset S2 -Full list of differentially expressed genes between resting and 2 h stimulated (no chemokine) conditions.
• Dataset S3 -Full list of GO terms significantly associated with differentially expressed genes between resting and 2 h stimulated (no chemokine) conditions.
• Dataset S4 -Full list of GO terms significantly associated with differentially expressed genes shared and non-shared between Jurkat and primary memory CD4 + T cells stimulated for 2 h.
• Dataset S5 -Full list of GO terms significantly associated with differentially expressed genes shared and non-shared between primary total CD4 + and primary memory CD4 + T cells stimulated for 24 h.
• Dataset S6 -Full list of GO terms significantly associated with genes expressed under resting conditions that are unique to Jurkat or total primary CD4 + T cells.

Open Peer Review
In the revised manuscript Felce et al. have added new analysis comparing resting Jurkat cell transcriptomic data to publicly available single cell sequencing data for CD4 T cells. This data rounds out the study well, and adds to the robustness of this analysis. It would be helpful to see the activated Jurkat profile included in this comparison to see whether activated Jurkat cells might even more closely align with proliferating CD4+ T cells.
As a minor criticism, it would be preferable to see gene names used in Figure 1E to enhance the readability of this sub-figure.
Overall, this study remains a helpful and relatively robust analysis of Jurkat transcriptomics, providing good molecular evidence to support the dogma within the immunology field that Jurkat cells can serve as a model of early T cell receptor signaling and activation but poorly recapitulate the complexities of later T cell activation, differentiation, and function.

Is the study design appropriate and is the work technically sound? Yes
Are sufficient details of methods and analysis provided to allow replication by others?
cells and memory CD4+ T cells. Authors then confirm divergence in a small number of surface markers using flow analysis of activated Jurkat or primary CD4+ T cells at various timepoints.
The main findings of the paper in comparing Jurkat dataset with that of stimulated memory CD4 or total CD4 cells (for 24h) are valid and interesting. Broadly, these findings would be consistent with the view in the field that Jurkat cells are capable of activating some components of the early T cell activation pathway in response to CD3/CD28 stimulation, but Jurkat activation is radically less complex than that seen in genuine T cells. The dataset generated here provides strong data to support that view, and contributes significantly to literature on Jurkat cells.
Critiques: Some of the specific choices made for activation seem quite arbitrary. The justification for choosing 2 hours post stimulation seems somewhat weak for instance. It would have been interesting to see more timepoints investigated via RNA-seq. 1.
The addition of chemokine stimulation in initial Jurkat stimulation seems to result in limited insight was gained through the addition of chemokine stimulation and no comparator primary T cell chemokine-stimulated dataset was investigated. Authors should either further investigate to compare primary and Jurkat data with chemokines and discuss or move chemokine data to supplemental data to streamline and focus the manuscript on key findings.

2.
The validity of CD4+ memory cells as a comparator of Jurkat cells is questionable. This specific memory CD4+ subset may reflect a very specific subtype of T cells. While there may be limited datasets available for RNA-sequencing, could some comparison not be made with the plethora of microarray data available for activated T cells? (see Immgen.org).
Investigating comparison with single cell RNA-seq to identify which T cell type most closely aligns with Jurkat cells might be an interesting question to probe as well.

3.
Overall, I see no major scientific oversights in this publication. The data is novel and interesting and this reviewer would recommend that the manuscript should be accepted with minor revision to address concerns discussed above.

Are sufficient details of methods and analysis provided to allow replication by others? Yes
If applicable, is the statistical analysis and its interpretation appropriate? I cannot comment. A qualified statistician is required.
Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results?

Yes
Competing Interests: No competing interests were disclosed.
I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Author Response 06 Jun 2021 James Felce, University of Oxford, Oxford, UK We thank the reviewer for their detailed assessment of our work and useful feedback. We have made the following amendments in relation to the following specific points: Some of the specific choices made for activation seem quite arbitrary. The justification for choosing 2 hours post stimulation seems somewhat weak for instance. It would have been interesting to see more timepoints investigated via RNA-seq.
We agree that it would be interesting to undertake a more thorough investigation of the kinetics of Jurkat transcriptional responses relative to primary cells. The original justification for the chosen two-hour timepoint was due to the need to focus on very early transcriptional responses in order to inform other work that was being undertaken to examine the architecture of the immunological synapse and how transcriptional changes may influence this. Unfortunately, collecting further RNA-Seq data at later timepoints is not possible within the scope of this project as the original funding for the project has been exhausted and the key authors have moved positions. We nonetheless think that the data collected at this early timepoint is still of use to the field.
The addition of chemokine stimulation in initial Jurkat stimulation seems to result in limited insight was gained through the addition of chemokine stimulation and no comparator primary T cell chemokine-stimulated dataset was investigated. Authors should either further investigate to compare primary and Jurkat data with chemokines and discuss or move chemokine data to supplemental data to streamline and focus the manuscript on key findings. This is a valid point. We have included these data primarily to act as a resource for others considering similar work, even though they do not advance the key conclusions of the study. Unfortunately the format of WOR does not permit supplemental material otherwise we would move these data out of the main manuscript. Even though they do not provide substantial insight, we believe that these data are worthy of inclusion as negative results. We have unfortunately not been able to find suitable public microarray data that match the activation conditions in human cells. The publications that we have found with similar conditions do not appear to have publicly deposited raw data (possibly due to their age), and the published gene lists are not in formats that we can process. However, please see the point below regarding single cell RNA-Seq as an alternative to determine the relative difference between Jurkats and different primary CD4 + subtypes.
Investigating comparison with single cell RNA-seq to identify which T cell type most closely aligns with Jurkat cells might be an interesting question to probe as well.
We thank the reviewer for this interesting suggestion. We have now included a comparison between our resting Jurkat data and public single cell RNA-Seq data of different primary CD4 + subsets, including both effector and central memory cells. This analysis revealed that the primary cells were broadly comparable in their difference from the Jurkat transcriptome, with the exception of proliferating CD4 + cells. This is presumably due to the activation of genes involved in cell cycle progression in both Jurkats and proliferating primary cells. The differences between primary cell populations was much smaller than those between primary and Jurkat, further reinforcing the notion that Jurkats possess a highly abnormal transcriptional profile. The fact that the difference between the primary CD4 + subtypes is small compared to the difference between primary CD4 + and Jurkats validates the use of memory CD4 + T cell subsets in comparisons.
demonstrate how reasonable is the use of Jurkat leukemic lymphoblastic cell line as the model to study the process of T cell activation in response to T cell receptor (TCR) engagement. The work was carried out by 4 researchers from UK working in the field of T cell biology and immunological synapsis. The question was addressed previously (for reference, see for example Bartelt et al., 2009) 1 where the differences in signalling events were revealed between leukemic cell lines and healthy activated cells from peripheral blood. The authors of the present study addressed an important and relevant questions about differences in the processes of activation in Jurkat and healthy primary T cells at transcriptional level. They present whole-transcriptome RNA-Sequencing data for Jurkat E6.1 cells in the resting state and during early activation response (antiCD3/CD28), in the presence and absence of the chemokines CXCL12 and CCL19, and also perform a basic comparison between transcriptional responses in Jurkat E6.1 cells and those in primary human T lymphocytes. The authors concluded that Jurkat E6.1 cells represent a simplified model of early T cell transcriptional responses.
I consider that this work is highly useful for all researchers working in the field of T cell biology, both with healthy T cells and leukaemia of T lineage. I have some observations regarding article presentation: Taking into account that behaviour characteristics (proliferation rate, population doubling time, etc) and the pattern of gene expression may depend on passage number, the corresponding information (range of passage numbers) should be indicated in the Materials and methods section, to enable other researchers to reproduce the data of this study.

1.
It should be noted here that leukemic cell lines are frequently used not as general T cell model but rather as T-ALL model, in particular in the field of drug and drug targets discovery, etc. However, being derived from different patients and representing transformed T cells at different maturation stages, with different phenotypes and mutations patterns, neither Jurkat nor any other leukemic T cell line could be considered as a typical leukemic model. Therefore, comparative characterization of leukemic cell lines is emergently necessary. Instead of use of "gold standard" or "typical" leukemic cell line (Jurkat, or any other cell line) in these kind of studies, the use of several cell lines should be strongly recommended. Although optional, I would recommend to mention and clarify this point in Discussion section.
Please provide an explanation for your result that Jurkat-restricted genes did associate with "neurological biological processes".
Many of the Jurkat-enriched genes related to the neuronal-associated GO terms are also linked with several other GO terms, and are not highly neurone-specific. These include many genes involved in broad signal transduction mechanisms -including semaphorins, Wnt molecules, GPCRs, and Ephrins/Ephrin receptors. Many of these are known to be involved in immune processes (e.g. EPHB2: https://www.frontiersin.org/articles/10.3389/fimmu.2019.01473/full). Similarly, a number of transcription factors are also found in this dataset, and several are known to be expressed in T cells (e.g. FOXO6: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3053075/). The association of Jurkat-restricted genes with these terms almost certainly does not represent neuronal-specific differentiation of the cells, but rather that these genes have broad functionality and are also annotated as involved in neurological processes alongside their immune functions.

Competing Interests:
No competing interests were disclosed.