Embigin Is Highly Expressed on CD4+ and CD8+ T Cells but Is Dispensable for Several T Cell Effector Responses

Abstract T cell immunity, including CD4+ and CD8+ T cell immunity, is critical to host immune responses to infection. Transcriptomic analyses of both CD4+ and CD8+ T cells of C57BL/6 mice show high expression the gene encoding embigin, Emb, which encodes a transmembrane glycoprotein. Moreover, we found that lung CD4+ Th17 tissue-resident memory T cells of C57BL/6 mice also express high levels of Emb. However, deletion of Emb in αβ T cells of C57BL/6 mice revealed that Emb is dispensable for thymic T cell development, generation of lung Th17 tissue-resident memory T cells, tissue-resident memory T cell homing to the lung, experimental autoimmune encephalitis, as well as clearance of pulmonary viral or fungal infection. Thus, based on this study, embigin appears to play a minor role if any in αβ T cell development or αβ T cell effector functions in C57BL/6 mice.


INTRODUCTION
Cells in multicellular organisms are dispersed throughout; not all are in physical contact with one other.For cells to function, they must be able to respond to cues from their local microenvironment, distant hormones, as well as external stimuli.This intricate signaling network is facilitated by the involvement of four key types of adhesion molecules, namely cadherins, integrins, selectins, and members of the Ig superfamily (IgSF) (1).The IgSF recognizes additional IgSF members as well as integrins as their ligands.Such heterophilic interactions between endothelial cells and selectinsfound on leukocytes, platelets, and additional endothelial cellsare integral for cell integrity, diapedesis, and thrombus formation (2).Furthermore, these adhesion molecules play a vital role in signal transduction, forming a dynamic interface that governs cellular responses and interactions.
Embigin (encoded by Emb) is a transmembrane glycoprotein within the IgSF that has two Ig domains (35).It has been shown to be expressed in sebocyte progenitors in the sebaceous gland and provides structural support of these cells through binding to fibronectin (6).Studies have linked embigin to pancreatic adenocarcinoma and breast cancer, with both organs containing sebaceous glands (7,8).Additionally, it has been found to play a role in early rat prostate and mammary gland development (5), further suggesting its potential epithelial nicheinteracting factor.
It has been reported that embigin assists in the transportation of nutrients by acting as a chaperone for monocarboxylate transporters (MCTs) (9).These MCTs are necessary for the bidirectional flow of monocarboxylatessuch as lactate, b-hydroxybutyrate, and acetoacetateacross the plasma membrane (10).Cells that use glycolysis as a major source of energy, such as leukocytes, must remove lactic acid to prevent acidosis.Of the four members of the MCT family, MCT2 preferentially binds to its chaperone protein embigin (11).MCT2 is found in the spleen, heart, kidney, and brain and in leukocytes (12,13).Previous studies have demonstrated that inhibition of MCT1 impairs mouse T cell proliferation in vitro, although studies focusing on MCT2 function with regard to T cell function are limited (14).
Homeostasis of the hematopoietic stem and progenitor cell (HSPC) microenvironment is based on the adhesive interactions of HSPCs with nearby cells, hormones, growth factors, and components of the extracellular matrix (ECM).These adhesive interactions influence HSPC self-renewal and proliferative potential.Embigin is one of the components that makes up the ECM in the HSPC microenvironment (15).Recent studies have revealed that embigin serves a role in the regulation and retention of HSPCs via regulation of HSPC quiescence (16).Emb transcripts are highly expressed in the brain, visceral yolk sac, and foregut during mouse embryogenesis (3).
Cell adhesion molecules play an important role in cell attachment, migration, and metastases.Integrins are known to bind to and activate MMP2, leading to the breakdown of the ECM; this plays a pivotal role in angiogenesis.Furthermore, integrin regulates cell attachment, spreading, and migration, playing a role in metastasis.In endothelial cells, integrin prevents apoptosis through the intrinsic apoptosis pathway (17).With integrin as its ligand, one could postulate that embigin is associated with malignancy.Emb has been found to be expressed in a variety of cancer cell types, including pancreatic and breast cancer.Embigin suppresses tumorigenesis in breast cancer cells while promoting pancreatic cancer progression (7,8).Homeobox C8 binds to the Emb promoter and inhibits Emb expression, which leads to an increase in proliferation, anchorage-independent growth, and migration of breast cancer cells (8).Emb expression is elevated in pancreatic ductal adenocarcinoma and involved in epithelial-to-mesenchymal transition in pancreatic cancer via the TGF-b signaling pathway (7).
Data from ImmGen (18) show that Emb is highly expressed in both ab and gd T cells.Prior work from our group showed that Emb is also highly expressed on lung CD4 1 tissue-resident memory T (TRM) cells (19) after immunization with the adjuvant Escherichia coli labile toxin A1 (LTA1) and outer membrane protein X (OmpX) from Klebsiella pneumoniae.Also, a recent study has demonstrated that Emb is highly expressed in skinresident NK cells (20).However, the function of Emb in T cells remains unclear.To study this, we generated Cd4cre × Emb fl/fl mice and used several CD4 1 T celldependent models to phenotype these mice, including lung TRM cell generation by mucosal vaccination, the lung homing property of lung TRM cells, the experimental autoimmune encephalomyelitis (EAE) model, Pneumocystis infection, as well as generation of CD8 1 T cell memory after influenza infection.Despite loss of embigin protein, embigin was found to be dispensable for both generation of CD4 responses as well as generation of CD8 1 T cell memory.

Mice
Emb fl/fl mice were generated by Cyagen (Santa Clara, CA).Cd4-cre 1 mice (B6.Cg-Tg(Cd4-cre)1Cwi/BfluJ) and B6.SJL-Ptprc a Pepc b /BoyJ mice (CD45.1 1 C57BL/6) were obtained from The Jackson Laboratory.Male and female mice (612 wk old) were used for studies.All mice were housed and bred at the Tulane University Department of Comparative Medicine Facility.Animals were housed in a pathogen-free environment.All experiments were performed using sex-and age-matched controls and approved by the Institutional Animal Care and Use Committee of Tulane University.

Ag and adjuvant preparation for immunization
We have previously shown that immunization with OmpX and LTA1, a mucosal Th17 adjuvant, derived from the A1 domain of the heat-labile toxin from Escherichia coli, drives the production of OmpX-specific lung CD4 1 TRM cells (19,22).To investigate whether Emb is required in CD4 1 lung TRM cell generation, we performed LTA1/OmpX immunization with isofluraneanesthetized mice using the oropharyngeal aspirationtongue pull technique and boosted with the same vaccine 3 wk later.Emb fl/fl Cd4-cre À and Emb fl/fl Cd4-cre 1 mice (male, 610 wk old) were used for this experiment.
Experimental K. pneumoniae infection K. pneumoniae-396 (K1 strain) was prepared as previous reported (19).Briefly, K. pneumoniae-396 (K1 strain) was grown in 30 ml of tryptic soy broth (Difco) overnight at 37 C with shaking at 233 rpm.Cultures were then diluted at 1:100 and grown in the same conditions for 2.5 h to achieve early logarithmic phase.The concentration of K. pneumoniae was determined by measuring the OD at 600 nm.Bacteria were pelleted and washed twice in cold PBS and then resuspended in PBS to the desired concentration.Mice were infected with 1 × 10 4 CFU intratracheally and sacrificed at 24 h postinfection.The lungs and spleens were homogenized and diluted in PBS, then plated on Luria-Bertani agar plates for CFU.

Single-cell suspension preparation
Single-cell suspensions of spinal cords were prepared as previously reported (23).Spinal cords were isolated and placed in icecold RPMI 1640 medium containing 27% Percoll, and pressed through a 70-mm cell strainer (Fisher Scientific).The resulting cell suspension was brought to a volume of 50 ml with 27% Percoll, mixed, and centrifuged at 300 × g for 15 min.The pellet was kept on ice, while the myelin layer and the supernatant were transferred to a new 50-ml tube, homogenized by shaking, and centrifuged again at 300 × g for 15 min.The pellets were then combined and washed three times in RPMI 1640 medium at 4 C. Single-cell suspensions of spleens were prepared by being pressed through a 70-mm cell strainer (Fisher Scientific) and centrifuged at 300 × g for 5 min.ACK lysis buffer (Life Technologies) was added to the cell pellet for 4 min and washed then resuspended in complete IMDM medium.
The lungs were minced manually with dissection scissors and digested in 4 ml IMDM (Life Technologies) with 2 mg/ml collagenase (Sigma-Aldrich) and 80 U/ml DNase1 (Sigma-Aldrich) at 37 C for 1 h.Digested tissue was strained through a 70 mm cell strainer (Fisher).ACK (ammonium-chloridepotassium) lysis buffer (Life Technologies) was added to the cell pellet for 4 min and then washed and resuspended in complete IMDM medium.

IFN-g ELISPOT assays
ELISPOT assays were conducted in Millipore pore plates to evaluate the frequency of background and Ag-stimulated spot forming units using an ELISPOT Flex: mouse IFN-g (ALP) kit (no.3321-2A, Mabtech).Lung cells (10 5 per well) were stimulated with 2 mg/ml Pneumocystis in triplicate.

Intracellular cytokine staining and flow cytometry
Single-cell suspensions from mouse lungs and spleens were stimulated for 5 h with PMA (50 ng/ml; Sigma-Aldrich) and ionomycin (750 ng/ml; Sigma-Aldrich) and GolgiStop (1 mg/ml; BD Biosciences).Cells were then stained with Abs specific for surface markers, followed by permeabilization/fixation with Cytofix/Cytoperm (BD Biosciences) and stained with Abs against intracellular molecules.The cells were analyzed using a Cytek Aurora spectral flow cytometer.The influenza nuclear protein MHC class I (H-2D b /ASNENMETM) tetramer (conjugated to PE) was obtained from the National Institutes of Health Tetramer Core Facility (Emory University, Atlanta, GA).Staining was carried out at room temperature for 1 h in conjunction with other surface staining.

Influenza infection
Influenza A/PR/8/34 H1N1 was propagated in chicken eggs as previously described (26).Emb fl/fl Cd4-cre À and Emb fl/fl Cd4-cre 1 mice (male, 812 wk old) were infected with 30 plaque-forming units of influenza virus in 50 ml of sterile PBS intratracheally.Following infection, mice were monitored daily for weight loss for 7 d and survival.At 60 d postinfection, lung cells were harvested for flow cytometry.

Illumina mRNA library preparation
Prior to Illumina mRNA library construction, DNase-related total RNA was quantitated using the Qubit RNA HS (high-sensitivity) assay kit (no.Q32855, Thermo Fisher Scientific,).RNA quality (RNA integrity number) was determined on an Agilent TapeStation 4150 using Agilent RNA ScreenTape (no.5067-5576, Agilent), after which 0.3 mg of each total RNA (RNA integrity number > 8) was applied to generate mRNA libraries using an Illumina TruSeq stranded mRNA sample preparation kit (no.20020594, Illumina), following the Illumina TruSeq stranded mRNA sample preparation guide (Illumina document no.100000004049).Final cDNA libraries containing TruSeq RNA CD indexes (no.20019792, Illumina) were quantitated using a Qubit dsDNA HS (high-sensitivity) assay kit (no.Q32854, Thermo Fisher Scientific).The quality of the libraries was determined by running each on an Agilent TapeStation 4150 using an Agilent D1000 ScreenTape (no.5067-5582, Agilent).Smear analysis was performed using Agilent TapeStation software (version 4.1.1)with a range of 200600 bp to determine average size of each library.Size and concentration were then used to calculate the molarity of each library.All libraries were pooled at a final concentration of 750 pM with a spike-in of 2% PhiX Control v3 library (no.FC-110-3001, Illumina).Mixture of pooled libraries was loaded on an Illumina NextSeq P1 (300) reagent cartridge (no.20050264, Illumina).Paired-end and dual indexing sequence, 150 × 8 × 8 × 150, was performed on NextSeq2000, yielding $20 million paired-end reads per sample.Fastqs generated by Illumina BaseSpace DRAGEN analysis software (version 1.2.1) were applied for further data analyses.RNA sequencing data were deposited in Gene Expression Omnibus under accession number GSE253411 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE253411).

Statistical analysis
Statistical analysis was performed with Prism (GraphPad) software.For comparison between two groups, a Student t test was used.For analysis comparing three or more groups, we used one-way ANOVA with Tukey post hoc analysis.For analyses involving bacterial burdens, we performed a log transformation on the data and performed ANOVA on transformed data.A p value <0.05 was considered statistically significant.The p values are annotated as follows: *p # 0.05, **p # 0.01, ***p # 0.001, ****p # 0.0001.

RESULTS
Emb is highly expressed on CD4 1 T cells Previous work has shown that OmpX and LTA1 together elicit Ag-specific CD4 1 TRM cells in the lung, which can protect the lung against K. pneumoniae infection (19).Single-cell RNA sequencing analysis showed that the CD4 1 TRM cells highly expressed Emb. CD4 1 T cells from naive spleen also expressed Emb but to a lower extent (Fig. 1A).Data from ImmGen showed that Emb was also highly expressed on CD4 1 and CD8 1 T cells.The expression level of Emb on B cells was low (Fig. 1B) (21).However, there is currently no report about the function of Emb on CD4 1 and CD8 1 T cells.

Cd4cre × Emb fl/fl mice generation and validation
To study the function of Emb on T cells, we generated Emb fl/fl mice.The Emb gene is located on mouse chromosome 13.Nine exons were identified, with the ATG start codon in exon 1 and the TGA stop codon in exon 9. Exon 5 was selected as the conditional knockout (cKO) region.The KO of exon 5 resulted in a frameshift mutation.To engineer the targeting vector, homologous arms and the cKO region were generated by PCR using BAC (bacterial artificial chromosome) clone RP23-135K13 as a template.In the targeting vector, the Neo cassette was flanked by SDA (self-deletion anchor) sites.DTA (diphtheria toxin A) ImmunoHorizons was used for negative selection.C57BL/6N embryonic stem cells were used for gene targeting (Supplemental Fig. 1A).After Emb fl/fl mice were generated, they were crossed to Cd4-cre 1 mice to conditionally knockout Emb on CD4 1 T cells.Loss of embigin expression in lung TRM cells in Cd4cre × Emb fl/fl mice was validated immunizing mice with LTA1 and OmpX as previously described (19) and staining for surface embigin by flow cytometry (Supplemental Fig. 2).We found high levels of embigin expression in Cre À mice (WT), which is substantially reduced in Cre 1 mice (KO).Similar results were found on the CD4 1 and CD8 1 T cells from spleen and thymus (Supplemental Fig. 1B).

Emb is dispensable for T and B cell development
To study the function of Emb on T and B cell development, we harvested the spleen and thymus from 4-wk-old WT and Emb-deficient mice.In terms of the number of CD4 1 T cells, CD8a 1 T cells, CD4 À CD8a À T cells, CD4 1 CD8a 1 T cells and CD19 1 B cells, there were no significant differences between WT and KO mice in the spleen (Fig. 2A, 2C) and thymus  (A) WT and KO mice were immunized twice with 1 mg of OmpX and 10 mg of LTA1 intratracheally 3 wk apart.Four weeks after the prime immunization, the mice were challenged with 10 4 CFU of K. pneumoniae (K1 strain), and bacterial burden (CFU) in lungs and spleens was determined (Continued) (Fig. 2B, 2D).These results suggest that Emb deficiency does not affect T and B cell development.
Emb is dispensable for LTA1-induced CD4 1 TRM cells and vaccine efficacy It has been reported that the LTA1-induced CD4 1 T cells are TRM cells that highly express Cd44 and Cd69 (19).Because Emb is highly expressed in LTA1-induced CD4 1 TRM cells, we immunized and challenged WT and KO mice as previously described with LTA1/OmpX (19).Single cells from the mouse lung were generated and IL-17A and IFN-g production levels were validated by flow cytometry.In WT mice, $10% of CD4 1 T cells were IL-17Aproducing cells, which was comparable in KO mice (Fig. 3A).Furthermore, after Klebsiella challenge, both WT and KO mice showed equivalent protection in the lung and spleen compared with the naive control (Fig. 3B).Intracellular staining for IFN-g was validated on in vitrodifferentiated Th1 cells because the IFN-g 1 cell number is low in the lung CD4 1 T cells (Supplemental Fig. 3).To determine whether the Embdeficient CD4 1 T cells differ from WT CD4 1 T cells, we enriched the lung CD4 1 T cells from immunized WT and KO mice using magnetic beads.We performed bulk RNA sequencing on these cells and found the differentially expressed genes between WT and KO mice (Fig. 3C).Interestingly, Tlr2 and Il1a were elevated in Emb-deficient CD4 1 T cells, which indicated that the Tlr2 signaling pathway might be associated with Emb regulation.Embigin has been reported to function as a cell adhesion molecule and plays a role in cell migration (6,7).Also, we have previously shown that LTA1-induced CD4 1 TRM cells can home back to the lung after adoptive transfer (19).Therefore, we tested whether Emb deficiency affects the homing of adoptively transferred CD4 1 TRM cells.CD4 1 T cells were enriched by magnetic beads from the lungs of immunized WT or KO mice.The CD4 1 T cells were adoptively transferred to Cd45.1 1 mice i.v. as previously described (19).We observed no differences between WT and KO mice in terms of the number of CD45.2 1 cells found in the mouse lung by flow cytometry (Fig. 3D, 3E).Moreover, the frequency levels of IL-17Aproducing cells by ELISPOT (Fig. 3F) were similar, suggesting that Emb is dispensable of lung TRM cell homing in this model (Fig. 3E).

Emb is dispensable for EAE pathogenesis
To study the roles of Emb in the development of EAE, a disease model that requires Th17 cells, we immunized WT and KO mice with MOG 3555 peptide and monitored the disease by physical examination daily.There were no significant differences in clinical score (Fig. 4A).Also, there were no differences in the percent of IL-17A 1 cells, IFN-g 1 cells, and Tregs between WT and KO mice, indicating that Emb is dispensable for EAE pathogenesis (Fig. 4B, 4C).
T cell expression of Emb is dispensable for T cell priming and clearance of Pneumocystis infection Clearance of Pneumocystis infection in mouse and humans requires CD4 1 T cell immunity (27,28).Thus, we infected WT and KO mice with P. murina by oropharyngeal aspiration.At 2 and 6 wk postinfection, the lungs were removed and RNA was extracted to assess fungal burden by RT-qPCR.No differences were detected between WT and KO mice (Fig. 5A).Also, part of the lungs was harvested to generate single-cell suspensions for IFN-g ELISPOT under PC Ag stimulation.There were also no differences between WT and KO mice (Fig. 5B).These results suggest that Emb is dispensable for fungal CD4 1 T cell priming and fungal clearance.
Emb on CD8 1 cells does not play a role in the influenza PR8 model Emb is also highly expressed on CD8 1 cells.Therefore, we tested Cd4cre × Emb fl/fl mice in the influenza PR8 model, which is a CD8 1 celldependent model.WT and KO mice were infected with Influenza H1N1 PR8 at a dose of 30 PFU.The body weight and survival of the mice were monitored daily over time.Both WT and KO mice lost weight to 80% at day 7 postinfection in a similar trend (Fig. 6A).The probability of survival of WT and KO mice was the same (Fig. 6B).In addition, we used influenza H1N1 PR8vtetramer to detect the number of Ag-specific CD8 1 cells from the lung at 60 d postinfection.Still, no differences were found between WT and KO mice, which indicated that Emb may not play a role in the influenza PR8 model (Fig. 6C, 6D).

DISCUSSION
Embigin (encoded by Emb) is a type I transmembrane glycoprotein within the IgSF.It is known to play a role in embryonic endoderm development and HSPC regulation (3,16).However, few studies have been conducted investigating the role of embigin in lymphocytes.This study is (to our knowledge) the first to investigate the role of embigin in T cell development.Our results indicate that embigin does not play a significant role in T cell development.Importantly, however, note that hematopoietic stem cells emerge in the yolk sac and later migrate to fetal liver in the mouse, where primitive hematopoiesis takes place.During this stage, the initial wave of lymphopoiesis occurs, giving rise to the first lymphoid precursors (29).T cell development takes place in the thymus.The spleen then acts as a secondary lymphoid organ, where T cells mature after having completed their development in the thymus and bone marrow, respectively (29).This current study looked at T cell development in both the spleen and thymus.Although the results of this study indicate that embigin does not play a significant role in T cell development, future studies should be directed at  looking at Emb KO in mouse embryos to determine whether it plays a role in the earliest stages of lymphopoiesis.
Embigin has been shown to serve as a chaperone for MCT2, which is necessary for the transportation of nutrients into and out of the cell (9).Because both MCT2 and embigin expression levels are increased in lymphocytes, and effector T cells predominantly engage in aerobic glycolysis to meet their biosynthetic and metabolic demands, one could hypothesize that both MCT2 and embigin play an important role in maintaining T cell homeostasis by shunting lactate out of the cell (13,30).Previous studies have demonstrated that inhibition of MCT1 impairs mouse T cell proliferation in vitro, although studies on MCT2 function on T cell function are limited (14).
Based on our previous work, we identified a vaccine composed of the A1 domain of heat-labile toxin from E. coli (LTA1) adjuvant and OmpX from the K2 strain of K. pneumoniae, which can elicit serotype-independent protection against the K1 strain by generating lung CD4 1 TRM cells (19).By singlecell RNA sequencing, we found that the CD4 1 TRM cells as well as CD4 1 T cells from naive spleen highly expressed Emb. Additionally, CD8 1 T cells also highly expressed Emb.However, the function of embigin in ab T cells has not been described.By using Cd4cre × Emb fl/fl mice, we generated mice with homozygous deletion of Emb in both CD4 1 and CD8 1 T cells and used this line to test the role of embigin in several T celldependent models.In our study, deletion of Emb did not affect generation of CD4 1 and CD8 1 effector or memory responses in various T celldependent models.Our previous studies on lung CD4 1 TRM cells found that these CD4 1 TRM cells expressed several unique adhesion molecules such as embigin, and preferentially homed to lung (19).However, this study demonstrated that Emb was dispensable for the homing property of lung-derived CD4 1 Th17 TRM cells to the lung.
Our research also was (to our knowledge) the first to link embigin to Tlr2 and Il1a signaling.Both Tlr2 and Il1a were elevated in Emb-deficient CD4 1 T cells, which indicated that the Tlr2 signaling pathway might be influenced by Emb expression.Bulk RNA sequencing of CD4 1 TRM cells from both WT and KO mice revealed that Tlr2 was elevated in Emb-deficient CD4 1 mice, suggesting that Emb might have an inhibitory effect on Tlr2 signaling.CD4 1 T cellintrinsic Tlr2 is known to help mediate Th17 development via the production of IL-6 and TGF-b (31,32).Furthermore, Tlr2 engagement with CD8 1 T cells leads to increased activation, proliferation and memory cell development (33).This could suggest that embigin, through

FIGURE 2 .
FIGURE 2. T and B cell development in Cd4cre × Emb fl/fl mice.(A) Representative dot plots of spleen CD4 1 T cells, CD8a 1 T cells, CD4 À CD8a À T cells, CD4 1 CD8a 1 T cells, and CD19 1 B cells in Emb fl/fl CD4-cre À (WT) and Emb fl/fl CD4-cre 1 (KO) mice.(B) Representative dot plots of spleen CD4 1 T cells, CD8a 1 T cells, CD4 À CD8a À T cells, and CD4 1 CD8a 1 T cells in WT and KO mice.(C and D) Quantitative data of (A) and (B) (n 5 3 per group from one experiment).Data are presented as means 6 SEM.Significant differences were calculated with an unpaired Student t test.

(
A) WT and KO mice were immunized for EAE with MOG 35-55 peptide, as described in Materials and Methods, and EAE score is shown (n 5 5 per group, representative from two independent experiments).(B) Representative dot plots of spleen and spinal cord IL-17a 1 cells, IFN-g 1 cells, and CD4 1 Foxp3 1 cells in WT and KO mice.were euthanized at day 30 after immunization.(C) Quantitative data of (B).Data are presented as means 6 SEM.Significant differences were calculated with an unpaired Student t test.

FIGURE 5 .
FIGURE 5. Embigin on CD4 + T cells is dispensable for T cell priming and clearance of Pneumocystis infection.

(
A) WT and KO mice were infected with 100 ml of P. murina inoculum (2 × 10 5 cysts) via oral pharyngeal administration.Two weeks (left) (n 5 3 per group from one experiment) or 6 wk (right) (n 5 5 per group from one experiment) later, lungs from all mice were removed for RNA isolation and fungal burden assessment by RT-qPCR.(B) Two weeks (left) or 6 wk (right) later, single cells from mouse lungs were prepared for IFN-g ELISPOT.Data are presented as means 6 SEM.Significant differences were calculated with an unpaired Student t test or one-way ANOVA followed by a Tukey multiple comparison test.****p # 0.0001.