Expression of Janus kinase 3 in human endothelial and other non-lymphoid and non-myeloid cells.

Members of the Janus kinase (Jak) family of protein tyrosine kinases have recently been implicated in the proximal signal transduction events of cytokine receptors. Jak3, a newly discovered member of this family, is believed to be normally limited in its expression to cells of the lymphoid and myeloid lineages. Herein we show that Jak3 is expressed in primary human vascular cells, as well as other non-lymphoid and non-myeloid cell types. Reverse transcriptase-polymerase chain reaction and Northern blot analysis revealed that Jak3 mRNA was expressed at low levels in human umbilical vein endothelial cells (HUVEC), human aortic smooth muscle cells (HASMC), A549 (human lung carcinoma), and DLD-1 (human colon adenocarcinoma) cells. Higher basal levels of Jak3 mRNA were detected in HMEC-1 (human microvascular cell line) and HepG2 (human hepatocellular carcinoma) cells. Jak3 mRNA expression was induced in HUVEC, HMEC-1, and HASMC by treatment with interleukin-1beta, tumor necrosis factor-alpha, interferon-gamma, and lipopolysaccharide. Jak3 protein was detectable at low levels in untreated HMEC-1, and these levels increased significantly with cytokine treatment. Furthermore, Jak3 protein was phosphorylated upon treatment of these cells with interleukin-4. This work shows that Jak3 is expressed or inducible in human vascular endothelial, vascular smooth muscle, and other non-lymphoid and non-myeloid cells, suggesting a broader role for Jak3 in the cytokine signal transduction of these cells.

normally in transformed hematopoietic and epithelial cell lines, as well as breast cancer tissue (10,11). Physiologic Jak3 expression, however, has not been shown to occur in primary cells or tissues outside of the lymphoid or myeloid lineages.
Jak3 is involved in the signal transduction of the family of cytokine receptors that utilize the common chain (␥c) of the IL-2 1 receptor (IL-2R), including IL-2R, IL-4R, IL-7R, IL-9R, and IL-15R (12). Jak3 associates with the ␥c chain and is activated in response to ligand binding (13,14). A variety of non-lymphoid, non-myeloid cell types respond to IL-4, with vascular tissue, especially vascular endothelial cells, being profoundly affected (15). For example, vascular cell adhesion molecule-1 (VCAM-1), interleukin-6 (IL-6), and monocyte chemotactic protein-1 are expressed in endothelium in response to IL-4 (16 -18). Given that endothelium is IL-4 responsive, that Jak3 binds the ␥c chain, and that Jak3 has not been reported to be expressed in vascular endothelium, we sought to determine whether endothelial cells express Jak3, and if so, whether they utilize Jak3 in IL-4 receptor signaling. Herein we provide the first evidence of primary human cells outside of the lymphoid and myeloid lineages expressing Jak3. We show that primary human endothelial and vascular smooth muscle cells, as well as a variety of non-lymphoid and non-myeloid cell lines, have detectable, albeit low, basal levels of Jak3. Jak3 levels are significantly increased in vascular endothelial and smooth muscle cells by stimulation with IL-1␤, TNF-␣, IFN-␥, and LPS. Finally, IL-4 treatment of the microvascular endothelial cell line HMEC-1 induces the phosphorylation of Jak3, thus suggesting a role of Jak3 in the signal transduction cascade of IL-4, and potentially other members of the ␥c chain receptor family, in cells outside of the myeloid and lymphoid lineages.
Detection of Jak3 in Human Endothelial and Other Non-lymphoid and Non-myeloid Cells by RT/PCR-Total RNA was prepared using the method of Chomczynski and Sacchi (19), and cDNA was prepared using Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.). Initial detection and cloning of the 0.64-kb Jak3 amplicon from HUVEC cDNA was performed by PCR (Taq polymerase, Life Technologies, Inc.) using a sense primer with a NotI cloning site linker (5Ј-gagagcggccgcaggcaagttgcactcatggcacctccaagt-3Ј), and an antisense primer with a ClaI cloning site linker (5Ј-gagaatcgattcacgaagctcaggccctggatcaggtcgc-3Ј). PCR conditions were as follows: 93°C for 3 min, then 35 cyles of 93°C for 60 s, 60°C for 60 s, and 72°C for 120 s, and finally one cycle at 72°C for 3 min. The 0.64-kb Jak3 amplicon was digested with NotI and ClaI, subcloned into the Bluescript pBSSKII vector (Stratagene, La Jolla, CA), and sequenced (Sequenase, Amersham). The amino acid sequence was predicted from the nucleic acid sequence and was compared to published sequences, using the Gene Works sequence analysis program.
For RT/PCR detection of Jak3 transcripts, cDNA was prepared from a variety of cells, and Jak3 target sequence was amplified as described above. The PCR amplicons were size-fractionated on a 1% agarose gel, which was then dried. The gel was then hybridized with a mixture of two primers specific for Jak3 and internal to the primers used for the initial amplification. The primers were labeled with T4 polynucleotide kinase and [␥-32 P]ATP. Hybridization and washing of the dried gel was done as described previously (20). The sequences of the internal primers are as follows: 5Ј-atccctcagcgttcatgcagcc-3Ј and 5Ј-cactcaccctgctccttgagactgagg-3Ј. For the detection of ␤-actin by PCR, the above PCR conditions were used on parallel cDNA samples with the following ␤-actin specific primers: 5Ј-acgatggaggggccggactc-3Ј and 5Ј-atggatgatgatatcgccgcg-3Ј. The gel was then stained with ethidium bromide and photographed.
Northern Blot Analysis-Forty micrograms of total RNA from each sample was run on a 1% denaturing agarose gel, and blotted onto a Zeta-Probe membrane (Bio-Rad). A 3.4-kb cDNA containing the entire Jak3 coding region was obtained from HUVEC cDNA by using long range PCR (Elongase, Life Technologies, Inc.) with the following primers containing NotI and ClaI cloning site linkers: 5Ј-gagagcggccgcaggcaagttgcactcatggcacctccaagt-3Ј, 5Ј-gagaatcgatcgggcaggagctatgaaaaggacagggagt-3Ј. PCR conditions were as follows: 93°C for 30 s, then 35 cycles of 93°C for 30 s and 68°C for 10 min. A 3.4-kb band was isolated and cloned into the Bluescript pBSSKII vector, and shown to be Jak3 by restriction mapping and by PCR using multiple internal primer pairs (data not shown). Membranes were hybridized with the 32 P-labeled (Random Primed Labeling Kit, Boehringer Mannheim) 3.4-kb Jak3 probe overnight at 50°C in 50% formamide, 10 ϫ Denhardt's reagent, 2% SDS, and 5 ϫ SSPE. They were then washed once with 2 ϫ SSC at room temperature for 30 min, once with 2 ϫ SSC at 50°C for 30 min, once with 0.2 ϫ SSC at 65°C for 30 min, and exposed for 5 days at Ϫ70°C. Membranes were then stripped in boiling water containing 0.1% SDS for 10 min, and rehybridized with the following ␤-actin specific primer end-labeled with T4 polynucleotide kinase and [␥-32 P]ATP as described previously: 5Ј-acgatggaggggccggactc-3Ј. All hybridization and wash conditions were the same as above, except the 65°C wash step was omitted.
Western Blot Analysis-Thirty to fifty million cells were lysed at 4°C for 20 min in 1 ml of buffer consisting of 50 mM Tris-HCl (pH 7.5), 1% Triton X-100, 150 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 5 mM iodoacetamide, 10 g/ml leupeptin, 10 g/ml aprotinin, 1 mM sodium orthovanadate, and 1 mM EDTA. Lysates were cleared of nuclei by centrifugation. One microgram of Jak2 or Jak3 antisera (Santa Cruz Biotechnology Inc., Santa Cruz, CA) and 10 l of protein A-agarose were added, samples were rotated at 4°C for 1-8 h, and the protein Aagarose collected by centrifugation and washed four times in lysis buffer. Samples were run on an 8% SDS-polyacrylamide electrophoresis gel, then blotted onto nitrocellulose.
For the Jak3 and Jak2 immunoblots, membranes were blocked with 4% dry milk overnight at 4°C, then incubated with 1 g/ml Jak2 or Jak3 antisera in Tris-buffered saline and 0.05% Tween 20 (TBS/Tween) for 1 h at room temperature. The membranes were washed in TBS/ Tween for 30 min, and then incubated for 30 min with a 1:7500 dilution of a horseradish peroxidase-linked goat anti-rabbit antibody (U. S. Biochemical Corp.). Membranes were washed again for 30 min, and developed by enhanced chemiluminescence (ECL, Amersham).
For phosphotyrosine immunoblots, membranes were blocked with 4% dry milk and 3.5% bovine serum albumin for 4 h at room temperature, and then incubated with 1 g/ml biotinylated 4G10 (gift of R. Schreiber) in TBS and 0.2% Tween 20, 2% goat serum, 0.5% bovine serum albumin, and 1% fish gelatin at room temperature for 1 h. The blot was washed for 30 min at room temperature with TBS/Tween, incubated with streptavidin-horseradish peroxidase (1:1500, Life Technologies, Inc.) for 30 min, washed again for 30 min, and developed by ECL. Membranes were then stripped and reprobed with Jak3 antisera as described above.
Densitometry-To quantify the levels of Jak3 protein and Jak3 phosphorylation, Western blot films were analyzed using a LKB Bromma Enhanced Laser Densitometer, and the area under the curve was used to calculate relative protein levels.

Detection of Jak3 in Human Vascular Endothelial Cells-
Initial RT/PCR screening of cDNA obtained from HUVEC with Jak3 specific primers revealed a 0.64-kb amplicon, as predicted by the cDNA sequence (data not shown) (7). The HUVEC amplicon was subcloned, sequenced, and shown to be essentially identical to the published sequence for Jak3, differing in only two places (Fig. 1). A 3-base pair deletion resulted in a loss of serine at position 112, and a single nucleotide change at position 152 replaced threonine with alanine. Further PCR screening with human Jak3-specific primers showed Jak3 expression in the monocytic cell line U937, the T cell line MT2, HASMC, A549, and HMEC-1 cells (data not shown). PCR analysis using analogous murine Jak3-specific primers showed Jak3 expression in the murine fibroblast cell line L929, and in the murine macrophage cell line RAW 264.7 (data not shown).
Since RT/PCR screening showed the presence of Jak3 mRNA, we wanted to determine if Jak3 protein was detectable. HMEC-1 and Jurkat cell lysates were immunoprecipitated with Jak3 and Jak2 specific antisera, and immunoblotted for Jak3 ( Fig. 2A). A band of approximately 120 kDa was strongly detected in Jurkat cell lysates. A less intense band of similar size was detected in HMEC-1 lysates, indicating lower levels of Jak3 expression relative to Jurkat cells. No Jak3 specific band was seen in the Jak2 immunoprecipitates from HMEC-1 cells, nor in immunoprecipitates using Jak3 antisera saturated with specific peptide, thus establishing the specificity of the Jak3 antisera. A faint band was noted in the Jak2 immunoprecipitates from Jurkat cells that corresponded in size to Jak3 protein, which may represent a slight cross-reactivity of Jak2 antisera. Western blots were stripped and reblotted with Jak2specific antisera, showing a strong Jak2 band in the Jurkat and HMEC-1 cell lines (Fig. 2B). The Jak2 band ran slightly higher than the Jak3 band, as expected by their sizes. A faint band corresponding to Jak3 protein was detected in Jurkat cells immunoprecipitated with Jak3-specific antisera upon immunoblotting with Jak2 antisera. This is also consistent with slight cross-reactivity of the Jak2 antisera.
Jak3 Is Induced upon Cytokine Treatment of Vascular Endothelial and Vascular Smooth Muscle Cells-Recently, the expression of Jak3 has been shown to be induced in human monocytes by treatment with LPS and IFN-␥ (9). Endothelium is responsive to LPS and IFN-␥, as well as to the inflammatory cytokines IL-1␤ and TNF-␣. We therefore treated a variety of non-lymphoid and non-myeloid cell types with a cytokine mixture consisting of IL-1␤, TNF-␣, IFN-␥, and LPS to determine the inducibility of Jak3 mRNA by RT/PCR analysis (Fig. 3A). Jak3 mRNA was detected at low levels in untreated HUVEC, HASMC, A549 cells, and the human colon adenocarcinoma line DLD-1. Higher basal levels of Jak3 mRNA were present in untreated HMEC-1, HepG2, and Jurkat cells. Jak3 expression was profoundly increased in HUVEC, HMEC-1, and HASMC following an 18-h incubation of the cells with the cytokine mixture. Treatment of A549, DLD-1, and HepG2 cells with the same mixture did not induce further expression of Jak3 mRNA. PCR controls using ␤-actin-specific primers confirmed that equivalent amounts of cDNA were used in all reactions (Fig.  3B). PCR reactions without added cDNA showed no Jak3 or ␤-actin specific bands (Fig. 3).
To further support our RT/PCR findings, we obtained a fulllength human Jak3 cDNA by PCR, and used it to probe Northern blots of total RNA isolated from the same cell types (

FIG. 2. Jak3 protein is present in human endothelial cells. A,
HMEC-1 and Jurkat cell lysates from fifty million cells were immunoprecipitated with Jak2 and Jak3 specific antisera, and then immunoblotted for Jak3. Specificity of the Jak3 antisera was confirmed in Jurkat lysates by preincubating the Jak3 antisera with specific peptide (GCETHAFTAHPEGKHHSLSFS) prior to Jak3 immunoprecipitation (Jak3ϩpep). B, the membrane in panel A was stripped and reblotted with Jak2 antisera. The Jak2 protein runs slightly higher than the Jak3 protein as predicted by their sizes. HASMC resulted in a significant induction of Jak3 mRNA. In contrast, stimulation of A549, DLD-1, and HepG2 cells did not increase Jak3 levels. Treatment of Jurkat cells with calcium ionophore A23187 and phorbol 12-myristate 13-acetate also resulted in an induction of Jak3 mRNA. The Northern blots were stripped and reprobed with a labeled ␤-actin-specific primer (Fig. 4B). The blots confirmed that there were equivalent amounts of RNA in all lanes, with the exception of phorbol 12-myristate 13-acetate-and A23187-treated Jurkat cells which had significantly less ␤-actin mRNA. This decrease appears to be due to a lower steady state level of ␤-actin mRNA in these cells, since subsequent hybridization of these Northern blots with a probe specific for hypoxanthine-guanine phosphoribosyl transferase and visualization of RNA on the membrane by ethidium bromide staining confirmed that similar amounts of RNA were loaded in all lanes (data not shown).

Jak3 Protein Is Induced by Cytokine Stimulation in Human Endothelial Cells, and Is Tyrosine Phosphorylated upon IL-4
Treatment-Since Jak3 mRNA levels in several non-lymphoid, non-myeloid cells are induced by treatment with a cytokine mixture, we wanted to determine if Jak3 protein levels were also increased. Furthermore, we wanted to determine whether Jak3 was functional in these cells. Because human endothelium is IL-4 responsive, we stimulated HMEC-1 cells for 24 h with the cytokine mixture, followed by treatment with IL-4 for 10 min. Jak3 was then immunoprecipitated from post-nuclear lysates, and immunoblotted with an anti-phosphotyrosine antibody. A protein of approximately 120 kDa was phosphorylated approximately 4-fold above basal levels as determined by scanning densitometry upon IL-4 treatment (Fig. 5A). This band was readily apparent in cytokine-pretreated cells, but barely detectable in cells without cytokine pretreatment. The membranes were stripped and reblotted with Jak3 specific antisera (Fig. 5B). Jak3 protein levels were induced approximately 8-fold in cytokine-stimulated human endothelial cells compared to unstimulated cells. Furthermore, the protein phosphorylated upon IL-4 treatment was indeed Jak3. The increased level of tyrosine-phosphorylated Jak3 in IL-4-treated cells was not due to increased Jak3 protein levels, since the level of Jak3 protein in IL-4-treated cells was equal to or slightly less than the level of Jak3 in cells treated with the cytokine mixture alone (Fig. 5B). DISCUSSION The specificity of tissue expression of Jak3 has been a matter of controversy. Our findings of the presence of Jak3 in vascular and other non-myeloid and non-lymphoid cells is consistent with the fact that a variety of cell types respond to IL-4 (15,21), and that the IL-4 receptor utilizes the ␥c chain and Jak3 (12). Also, Jak3 has been identified in human breast cancer cells and other epithelial cancer cell lines (10,11). However, physiologic Jak3 expression appeared to be predominately limited to lymphoid and myeloid cell types (7,8). Herein we show that a variety of non-lymphoid, non-myeloid human cells express Jak3. More importantly we show that primary human endothelial and vascular smooth muscle cells express Jak3, and that this expression can be induced in response to inflammatory cytokines. The preliminary reports on the normal expression pattern of Jak3 were not done with cytokine-stimulated tissues (7,8), and thus its expression in these tissues may have been missed. Furthermore, these studies only assessed Jak3 expression by Northern blot analysis. To detect significant levels of Jak3 mRNA in HMEC-1, HUVEC, and HASMC by Northern blot, it is necessary to treat cells with cytokines, as is the case in human monocytes (9). We were, however, able to detect Jak3 in a variety of cell types without stimulation by RT/PCR analysis (Fig. 3). Although our RT/PCR studies were not done in a semi-quantitative manner, the results directly paralleled the findings of Northern blot analysis (Fig. 4). This, together with the ␤-actin data, argue against the possibility that the higher Jak3 levels detected by RT/PCR in cytokinetreated cells resulted from varying quantity or quality of cDNA in the reactions. Furthermore, the RT/PCR analysis is useful in that it permits detection of Jak3 transcripts when Northern blot shows no apparent Jak3 transcripts. This added sensitivity is significant when assessing whether a particular cell type utilizes Jak3 for cytokine signaling, since a negative Northern blot does not rule out the possibility that Jak3 is expressed at low levels. Whether this low level of expression of Jak3 is sufficient for signaling remains to be seen. Also, the low level of expression of Jak3 in unstimulated endothelial cells also suggests why tyrosine-phosphorylated Jak3 was not detected in phosphotyrosine blots from HUVEC stimulated with IL-4 alone (22). We were able to detect the IL-4 enhancement of tyrosine phosphorylated Jak3 in HMEC-1 cells only after pretreating cells with a cytokine mixture. Our data suggest that human vascular cells, and possibly other non-myeloid and non-lymphoid cells can be induced to express significant levels of Jak3.
Vascular endothelial cells form the interface between blood and tissue, playing an important role in organ physiology and pathology. For example, expression of adhesion molecules and secretion of inflammatory proteins by endothelium is essential for the adhesion, activation, and transendothelial migration of immune cells (23). Cytokines are important regulators of endothelial cell function, with IL-4 having numerous positive and negative effects. IL-4 inhibits the production of the chemokine RANTES induced by TNF-␣ and IFN-␥ in endothelium, and inhibits the expression of tissue factor and the down-regulation FIG. 4. Jak3 mRNA is detectable in a variety of non-lymphoid  and non-myeloid cells. A, a 3.4-kb Jak3 cDNA was used to probe Northern blots from cytokine (cyto) untreated (Ϫ) or treated (ϩ) cells as described in the legend to Fig. 3. B, subsequent stripping and reprobing of the Northern blots in panel A with a labeled ␤-actin specific primer showed that equivalent amounts of RNA were loaded into each lane. The lower expression level in the induced Jurkat lane represents a down-regulation of ␤-actin, as similar levels of RNA were shown by ethidium bromide staining and hypoxanthine-guanine phosphoribosyl transferase expression (data not shown).
FIG. 5. Jak3 protein is inducible in human endothelial cells upon cytokine treatment, and tyrosine phosphorylated in response to IL-4. A, thirty five million HMEC-1 cells were untreated (Ϫ) or treated (ϩ) with the cytokine mixture (cyto) for 24 h, followed by no treatment (Ϫ) or treatment (ϩ) with 10 ng/ml IL-4 for 10 min. Cell lysates were then immunoprecipitated with Jak3 antisera, and phosphotyrosine (p-tyr) immunoblotting was performed. B, the membrane in panel A was then stripped and reprobed with Jak3 antisera. of thrombomodulin expression induced by TNF-␣, IL-1␤, and LPS in endothelium (24,25). IL-6 and monocyte chemotactic protein-1 are expressed in endothelium in response to IL-4 (16,17). VCAM-1 expression on endothelium is induced by IL-4, also synergizing with TNF-␣ in this regard (18). The function of Jak3 in these responses is unknown. Jak3 is present in unstimulated endothelium at very low levels, and it is unknown whether these levels are significant enough to allow for efficient signaling through the IL-4 receptor. It is possible that the forementioned synergistic actions of TNF-␣ and IL-4 results from the TNF-␣ induced production of Jak3, which then allows for a greater response to IL-4. In support of this, we have shown that TNF-␣ is sufficient to induce the expression of Jak3 in endothelium. 2 Recently, however, it was shown that the synergy between IL-4 and TNF-␣ in the expression of VCAM-1 on HUVEC occurs in the presence of protein synthesis inhibitors, thus arguing against the possibility that Jak3 synthesis was necessary for the synergy (18). Jak3 expression, however, was not directly assessed in this study, nor in any of the forementioned studies. What affect the induction of Jak3 has on these responses is a topic of further investigation.
Our data indicate that vascular cells can be induced to express Jak3 in vitro. Whether this response is seen in vivo is unknown. Furthermore, what role Jak3 may be playing in vascular physiology and pathobiology is also under investigation. VCAM-1 induction by IL-4 and TNF-␣, and the effects of Jak3 expression in this responses may be important to the pathogenesis of atherosclerosis. VCAM-1 binds to ␣4␤1 and ␣4␤7 integrins on circulating monocytes and lymphocytes (23). It has been postulated that this interaction is important for the recruitment of these cells to sites of inflammation and atherosclerotic lesions (26). We are currently determining if Jak3 is present in endothelium in atherosclerotic lesions, as well as in endothelium and other tissues during inflammatory responses.
Jak3 activation has been shown to be critical to the IL-4 induced proliferative response of the human premyeloid cell line TF-1 (27). Furthermore, endothelium has been shown to proliferate in response to IL-4 (28). Whether the induction of Jak3 in endothelium may affect its proliferative response to IL-4 remains to be seen. The ability of previously quiescent endothelium to proliferate is an essential component of the physiologic angiogenic response of wound repair, as well as the pathophysiologic angiogenic response in diseases such as rheumatoid arthritis, diabetic retinopathy, and solid tumor formation. The determination of Jak3 involvement in these responses will be of interest.
These results also show that the level of expression of Jak3 can vary in tumor cell lines (Figs. 3 and 4). Our data shows that Jak3 was present at significant levels in HepG2 cells, while it was detectable at low levels and uninducible in A549 and DLD-1 cells. These cell lines may be useful research tools in studying Jak3 signaling. Kinase-deficient cell lines have been used to study the role of other kinases in cytokine signaling.
For example, mutant cell lines were used to establish the requirement for Jak1 and Jak2 in the response to IFN-␣ and IFN-␥ (5).
It is clear that primary human vascular cells, as well as other cells outside of the lymphoid and myeloid lineages express or can be induced to express significant levels of Jak3. The function of Jak3 in these cells, and the new functions acquired upon its induction remain to be elucidated.