CCR5 and CXCR4) and Their Ligands by Human Hematopoietic Cell Lines M Majka, T Rozmyslowicz, M Honczarenko, J Ratajczak, MA Wasik, GN Gaulton and MZ Ratajczak

Home , CCR5, CXCR4, HIV, Jurkat cells

Leukemia (2000) 14, 1821–1832  2000 Macmillan Publishers Ltd All rights reserved 0887-6924/00 $15.00 www.nature.com/leu Biological signiﬁcance of the expression of HIV-related chemokine coreceptors (CCR5 and CXCR4) and their ligands by human hematopoietic cell lines M Majka, T Rozmyslowicz, M Honczarenko, J Ratajczak, MA Wasik, GN Gaulton and MZ Ratajczak

Department of Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA, USA

The aim of this study was to learn more about the role of the Chemokines may also exert different effects on hematopoiesis, HIV-related chemokine–chemokine receptor axes in human for example, they may regulate the adhesive properties of hematopoiesis. To address this issue we phenotyped 35 selec- cells,6 and/or influence trafficking and homing of cells in ted hematopoietic cell lines for the expression of CD4, CXCR4 27,28 and CCR5. We next evaluated the functionality of these chemo- various lympho–hematopoietic organs. kine receptors by calcium flux and chemotaxis assays, and by The aim of this study was to evaluate the expression of the the ability of SDF-1, MIP-1α, MIP-1β and RANTES to influence chemokine receptors on human cells derived from the differ- the growth of the cells expressing CXCR4 and/or CCR5. Lastly, ent hematopoietic lineages (erythroid, megakaryocytic, we examined whether human hematopoietic cell lines may myeloid, T and B lymphoid lines), and to learn more about secrete some HIV-related chemokines, and whether endogenously secreted chemokines might interfere with the infectability the biological role of chemokines in regulating the biology of of hematopoietic cells by X4 and R5 HIV strains. These results hematopoietic cells. demonstrate that: (1) HIV-related receptors are widely We observed that chemokine receptors are broadly expressed on human hematopoietic cell lines; (2) stimulation expressed on human hematopoietic cell lines, however, their of CXCR4 by SDF-1 induces calcium flux and chemotaxis in expression does not correlate uniformly with functional several hematopoietic cell lines more efficiently than stimu- capacity. Of note, we found that while SDF-1 activates cal- lation of CCR5 by receptor-specific β-chemokines; (3) chemo- + kines do not regulate proliferation of the hematopoietic cells; cium flux and chemotaxis in several CXCR4 human hematopoietic cell lines, MIP-1α, MIP-1β or RANTES did not stimu- and finally (4) infectability of the hematopoietic cells by HIV-1 + may be auto-modulated by endogenously secreted chemo- late calcium flux and chemotaxis of CCR5 cell lines kines. These data shed more light on the role of HIV-related employed in this study. Furthermore, chemokines did not chemokine–chemokine receptors axes in human hemato- influence the proliferation of CXCR4+ or CCR5+ cell lines, poiesis and interaction of hematopoietic cells with HIV. Leuke- even when the corresponding receptors were found to be mia (2000) 14, 1821–1832. Keywords: CXCR4; CCR5; chemokines; cell lines; chemotaxis; functional by calcium flux or chemotaxis assay. Our studies calcium flux also showed that human T and B hematopoietic cell lines may secrete HIV-related chemokines. We observed that these endogenously secreted chemokines may influence the infect- Introduction ability of hematopoietic cells by both R5 and X4 HIV strains. In toto these data are helpful in understanding and interpreting HIV viruses enter human cells after binding to CD4 antigen the role of HIV-related chemokine–chemokine receptor axes in addition to one of the chemokine receptors. R5 viruses in human hematopoiesis, and may provide important clues in (macrophagotropic) utilize the CCR5 co-receptor and X4 understanding interaction of hematopoietic cells with HIV. viruses (lymphotropic) the CXCR4 receptor, respectively. Some viral strains may also enter cells in a CD4-independent manner or utilize other chemokine receptors that function as Material and methods HIV-1 co-receptors for possible entry.1–7 Chemokine receptors possess a seven span transmembrane protein molecular struc- Cell lines ture and are coupled to G signaling proteins.8–11 The major R5 HIV-1 co-receptor (CCR5) binds the β-chemokines MIP- All cell lines employed in these studies were maintained in 1α, MIP-1β and RANTES, whereas the major X4 HIV co- RPMI medium (Gibco BRL, Long Island, NY, USA) sup- receptor (CXCR4) binds α-chemokine SDF-1.3–5 plemented with 10% bovine calf serum (BCS) (Hyclone, The role of the chemokine–chemokine receptor axes in reg- Logan, UT, USA). K-562, HL-60, HEL, Jurkat and UT-7 hema- ulating the biology of human hematopoietic cells is still not topoietic cell lines were purchased from ATCC (Rockville, fully understood. As an example, MIP-1α has been reported MD,USA). Other hematopoietic cell lines were maintained in to inhibit,12–15 stimulate,16,17 or have no effect on the prolifer- the laboratory of one of us (MAW) in agreement with the 18–20 guidelines for the characterization and publication of human ation of normal human hematopoietic cells. The other 29 two HIV-related CCR5 binding chemokines (MIP-1β and maligant hematopoietic cell lines. RANTES) did not affect the proliferation of cells.12,13 SDF-1 – ligand for the CXCR4 receptor has been reported to have no effect on the proliferation of human hematopoietic cells.21–25 FACS analysis of chemokine receptor expression In a recently published study, however, SDF-1 has been reported to inhibit growth of myeloid progenitor cell line.26 The expression of CCR5 and CXCR4 on different human hematopoietic cell lines was evaluated by FACS as described.7,25 The MoAbs employed for the chemokine receptor expressions were a gift from R&D Systems (Minneapolis, MN, USA). Correspondence: MZ Ratajczak, Department of Pathology & Labora- tory Medicine, University of Pennsylvania, 406A Stellar Chance Labs, Briefly, the cells were stained in PBS (Ca and Mg free) sup- 422 Curie Blvd, Philadelphia, PA 19104, USA; Fax: (215) 573-7049 plemented with 5% BCS (Hyclone). Primary MoAbs were Received 2 March 2000; accepted 25 May 2000 detected with secondary PE-conjugated goat anti-mouse Hematopoietic cell lines and HIV M Majka et al 1822

Figure 1 Expression of CXCR4, CCR5 and CD4 on human hematopoietic cell lines. Cells were stained with CXCR4, CCR5 or CD4 MoAb and subsequently detected with PE-goat anti-mouse MoAb. Panel a, selected myeloid cell lines. Panel b, selected lymphoid cell lines. Representa- tive data of triplicate experiments are shown.

Leukemia Hematopoietic cell lines and HIV M Majka et al 1823 MoAbs (Sigma, St Louis, MO, USA) (1:100). After the ﬁnal Isolation of mRNA from cell lines wash, cells were ﬁxed in 1% paraformaldehyde prior to FACS analysis. Analysis was performed using the FACscan (Becton Cells were lysed in 200 µl of RNAzol (Biotecx Labs, Houston, Dickinson, San Jose, CA, USA). TX, USA) +22 µl of chloroform as described7,25 and the aque- ous phase was collected and mixed with 1 volume of isopro- panol (Sigma). RNA was precipitated overnight at −20°C. The Chemotaxis studies RNA pellet was washed in 75% ethanol and resuspended in

three times-autoclaved H2O. All experiments were performed in triplicate. Briefly, after isolation the cells were resuspended in serum-free medium (106/ml) and equilibrated for 10 min at 37°C. Pre-warmed Reverse transcriptase-polymerase chain reaction (RT- serum-free medium (600 µl) containing the appropriate ligand PCR) (300 ng/ml) was added to the lower chamber of a Costar µ Transwell 24-well plate, 6.5 mm diameter, 5–8 M pore filter Briefly, mRNA (0.5 µg) was reverse-transcribed with 500 U of µ (Costar Corning, Cambridge, MA, USA). Subsequently, 100 l Moloney murine leukemia virus reverse transcriptase aliquots of the cell suspension were distributed to the upper (MoMLV-RT) and 50 pmol of an ODN primer complementary ° chambers and cultures were incubated at 37 C, 95% to the 39 of SDF-1 (59-CAC ATG TTG AAC CTC TTG TTT AAA humidity, 5% CO2 for 3 h. After 3 h of incubation the cells AGC-39). The resulting cDNA fragments were amplified using from the lower chambers were collected and the cell number 5UofThermus aquaticus (Taq) polymerase and primers spe- was scored by FACscan (Becton Dickinson). Briefly, the cells cific for the 59 end of SDF-1 (59-AAC GCC AAG GTC GTG were gated according to their forward scatter (FSC) and side- GTC GTG CTG-39). β-Actin mRNAs were amplified simul- ward scatter (SSC) parameters and counted during a 20 s taneously using specific primers as reported previously.7,25 acquisition at high flow rate. Data are demonstrated as a Amplified products (10 µl) were electrophoresed on a 2% aga- percentage of the input number of the cells. rose gel and transferred to a nylon filter and further docu- mented photographically. Specificity of the amplified products were further confirmed by Southern blotting (data not shown). Calcium flux studies by spectophotofluorimeter

Calcium flux studies on cells from different human hematopo- Infection of PB-1 cells with HIV ietic cell lines were performed using the spectophotofluorimeter as described.7,25 Briefly, cells were resuspended in pre- A T lymphocytic cell line PB-1 was maintained in RPMI warmed loading buffer (1 × Hank’s buffered salt solution con- (Gibco BRL) supplemented with 10% BCS (Hyclone).30,31 taining 1 mM CaCl ,1mM MgCl and 1% low IgG FBS) to a 2 2 Cells were infected at day 0 by BAL (R5 HIV) or HXB2 (X4 final concentration of 107/ml. Subsequently, Fura-2/AM HIV). HIV strains (a gift of Dr F Gonzales-Scarano, University (Molecular Probes, Eugene, OR, USA) was added to a final of Pennsylvania, Philadelphia, PA) were employed at a multi- concentration of 1–2 µM, and cells were incubated for 30 min plicity of infection (MOI) of 0.02. Briefly, 105 PB-1 cells were at 30°C. After incubation cells were washed once, resus- resuspended in 200 µl of medium containing 2 ng HIV per pended in loading buffer without FBS, stimulated with appro- infection, and subsequently 300 µl of conditioned medium priate ligand (500 ng/ml) and analyzed within 1 h. harvested from serum-free cultures of PB-1, HUT102B, 15A and K-562 cells were added to the cells. As controls we Proliferation assays employed medium alone and conditioned media that were boiled before addition to the cultures. The cells were subsequently incubated for 24 h in 37°C, 5% CO . After incu- Cells were resuspended in Iscove’s medium (Gibco BRL, 2 bation, the cells were washed out and resuspended in DMEM Grand Island, NY, USA) supplemented with 10% of the arti- + 10% BCS and placed into the tissue culture incubator. At ficial serum as previously described.30,31 Cultures were perfor- days 1, 5 and 10 of the culture medium aliquots for p24 ELISA med in 96-well plates. The cells from the various hematopo- assays were aspirated. ietic cells lines (2 × 103 cells/ml) were resuspended in 1 ml of the culture medium without or in the presence of 5 ng (low dose) or 50 ng/ml (high dose) of the following chemokines: SDF-1α, SDF-1β, MIP-1α, MIP-1β, RANTES (R&D, Minnea- Detection of p24 protein by ELISA polis, MN,USA). The cells in the wells were counted at days 7 and 14. Four wells were counted per point. Infection of the cells was measured by employing sensitive p24 ELISA (NEN, Boston, MA, USA). The p24 protein was detected in supernatants collected from HIV-infected PB-1 ELISA assays cells according to the manufacturer’s protocol.

Secretion of MIP-1α, MIP-1β and RANTES by hematopoietic cells was detected by Quantikine human MIP-1α, MIP-1β and Statistical analysis RANTES immunoassy (R&D) according to the manufacturer’s protocol. Cells were cultured for 24 h in serum-free medium Means and standard deviations were calculated on a Mac- and supernatants collected and subjected to the quantitative Intosh computer using Instat 1.14 (GraphPad, San Diego, CA, sandwich enzyme immunoassay. The sensitivity of the ELISA USA) software. Data were analyzed using the Student’s t-test assays for MIP-1α, MIP-1β and RANTES was .31 pg/ml, >31 for unpaired samples. Statistical signiﬁcance was deﬁned as pg/ml and .15 pg/ml, respectively. P , 0.05.

Leukemia Hematopoietic cell lines and HIV M Majka et al 1824 Table 1 Expression of HIV-related receptors on human myeloid Table 3 Expression of HIV-related receptors on human B lympho- cell lines cytic cell lines

Cells Lineage CD4 CCR5 CXCR4 Cells Lineage CD4 CCR5 CXCR4

TF-1 myelo +++ − − JIM-1 pro-B − − +++ KG-1 myelo −++−Nalm 6 pre-B − − +++ HL-60 myelo +++ − ++ 697 pre-B −++++ K-562 myelo −++−BL-1 B cells −±++ UT-7 myelo +++ + − 20A B cells −−+ HEL erythr ± +++ − A1 B cells −++− DAMI mega +++ + − 15A B cells −−+ Daudi B cells − +++ − −+++ Expression of CD4, CXCR4 and CCR5 was evaluated by employing HS Sultan B cells − +++ + FACS analysis. RS11846 B cells +++, .75% of cells positive; ++, .25–75% of cells positive; +, .5– 25% of cells positive; ±, ,5%; −, not detectable. Expression of CD4, CXCR4 and CCR5 was evaluated by employing FACS analysis. +++, .75% of cells positive; ++, .25–75% of cells positive; +, .5– ± , − Results 25% of cells positive; , 5%; , not detectable.

CD4 and major HIV chemokine co-receptors are HL-60 cells only. Interestingly, while CXCR4 was highly expressed on human hematopoietic cell lines expressed on more immature B lymphocytic cell lines (pro- and pre-B), CCR5 was expressed on more differentiated B cells We phenotyped 35 different human hematopoietic cell lines, (Table 3). In contrast, however, no similar correlation could including a series of myeloid, B and T cell lymphocytic cells, be made for CXCR4 and CCR5 expression and developmental for expression of CD4 and the major R5 HIV and X4 HIV-1 status of T cells (Table 2). strain-related chemokine co-receptors CCR5 and CXCR4, respectively. The data of representative FACS studies are shown in Figure 1, and summarized in Tables 1–4. We dem- Calcium flux in cells expressing CXCR4 and CCR5 onstrated that HIV-related chemokine co-receptors are widely co-receptors expressed on human hematopoietic cells. Interestingly, we found that the CD4 antigen was expressed We next examined whether cell lines that expressed CXCR4 on several human myeloid cells (TF-1, HL-60, HEL, UT-7 and (HL-60, NALM6, Jurkat, ATL2, L428, PB-1, 2A, BL-1, 697, Sul- DAMI) (Table 1) in addition to T lymphocytic cell lines (Table tan, 15A, Molt13 and Sez4) or CCR5 (K-562, HEL, DAMI, 2). Some T cell lines, however, did not express the surface DAUDI, JB-6, BL-1, 697, SULTAN, Sez4, PB-1, 2A and CD4 molecule (MOLT13, SUDHL-1, JB6 and 2A). As pre- Hs445) co-receptors on the surface would respond to chemo- dicted, all B lymphocytic cell lines were negative for CD4 kine binding in the calcium flux assay. expression (Table 3). Of note, one Hodgkin lymphoma cell Accordingly, target cells were loaded with Fura-2/AM and line, L540, strongly expressed the CD4 molecule on the sur- then stimulated with appropriate ligands – SDF-1 (Table 5) or face. MIP-1β (Table 6), respectively. We found that from the 13 cell Both major HIV-related chemokine co-receptors CXCR4 lines tested – HL-60, NALM6, Jurkat, ATL2, PB1, 2A and 697 and CCR5 were expressed on myeloid and lymphoid cells. cells showed calcium flux after stimulation by SDF-1 (Table The CCR5 co-receptor was expressed on five out of seven 6). An example of calcium flux in PB-1 and Jurkat cells after myeloid cell lines (Table 1), whereas CXCR4 was present on stimulation with SDF-1 is shown in Figure 2. Interestingly, we

Table 2 Expression of HIV-related receptors on human T lymphocytic cell lines

Cells Lineage (syndrome) CD4 CCR5 CXCR4

Jurkat immature T cells (T-ALL) ++ − +++** MOLT4 immature T cells (T-ALL) +++ − +++ MOLT13 immature T cells (T-ALL) − + +++ HUT102B mature T cells (ATLL) +++ − ++ ATL-2 mature T cells (ATLL) +++ − ++ C10MJ2 mature T cells (ATLL) +++− C91PL mature T cells (ATLL) +++ + +++ JB-6 mature T cells (ATLL) − +++ ± SUDHL-1 mature T cells (ATLL) −−++ Karpas 299 mature T cells (ATLL) +++ + − 2A mature T cells (ATLL) − ++ +++ Sez 4 mature T cells (SS) +++ ++ + PB-1 mature T cells (SS) +++ ++ +++

Expression of CD4, CXCR4 and CCR5 was evaluated by employing FACS analysis. T-ALL, T cell acute lymphocytic leukemia; ATLL, adult T cell leukemia; ALCL, anaplastic large cell lymphoma; SS, Sezary syndrome (leukemic form of cutaneous T cell lymphoma); +++, .75% of cells positive; ++, .25–75% of cells positive; +, .5–25% of cells positive; ±, ,5%; −, not detectable.

Leukemia Hematopoietic cell lines and HIV M Majka et al 1825 Table 4 Expression of HIV-related receptors on human Hodgkin’s To our surprise, none of the 12 CCR5+ cell lines employed disease (HD) and natural killer (NK) cell lines as determined by in this study showed calcium ﬂux after stimulation by MIP- FACS analysis 1β (Table 6). Similarly, none of the CCR5+ cell lines showed chemotaxis to MIP-1β or RANTES. Cells Lineage CD4 CCR5 CXCR4 Chemotactic responses of selected hematopoietic cell lines (PB-1, ATL-2, 2A and Jurkat cells) to α- and β-chemokines are L540 HD ++ ± ± + L428 HD − ++ +++* presented in Figure 3. All cells tested were CXCR4 (Table 3) Ryan HD −++−and showed chemotaxis to SDF-1 (Table 5). SDF-1, in parti- Hs445 HD −+++cular, was a very strong chemoatractant for Jurkat and PB-1 YT NK −+−cells (Figure 3). Both 2A and PB-1 cells, which also express CCR5, failed to show chemotaxis to MIP-1β (Figure 3, Table +++, .75% of cells positive; ++, .25–75% of cells positive; +, .5– 6). 25% of cells positive; ±, ,5%; −, not detectable.

+ Table 5 Biological responses of CXCR4 cell lines to stimulation Effects of MIP-1α, RANTES, MIP-1β and SDF-1α on by SDF-1 + + the proliferation of CCR5 and CXCR4 cell lines Cells Lineage Calcium flux Chemotaxis to SDF-1 SDF-1 The effect of chemokines on cell proliferation remains controversial.12–15,26 Therefore, we examined whether the CCR5 and HL-60 myeloid + +++ CXCR-4 receptor binding chemokines (MIP-1α, MIP-1β, +++ NALM6 B cell RANTES and SDF-1) affect proliferation of 10 CCR5+ and/or BL-1 B cell ND − + + +++CXCR4 cell lines. To address this issue we selected 10 CCR5 697 B cell + HS Sultan B cell ND + and/or CXCR4 cell lines as determined by FACS analysis. Of 15A B cell −−note, even where one particular receptor protein was not Sez 4 T cell ND − detectable by FACS on some of these cell lines (Tables 1–4) Jurkat T cell + +++ we were always able to detect its expression at mRNA level ++ ATL2 T cell by RT-PCR (data not shown). Subsequently, these 10 selected PB1 T cell + +++ + + MOLT13 T cell ND − CCR5 and/or CXCR4 cell lines (HEL, DAUDI, KG-1a, L-428, 2A T cell ++ATL-2, HL-60, PB-1, 2A, K-562 and Jurkat) were exposed to L428 HD −−MIP-1α, MIP-1β, RANTES or SDF-1 (Figure 4). This rep- resented four myeloid (HEL, KG-1a, HL-60 and K-562) and six +++, .20% of cells after 3 h show chemotaxis; ++, .10–20% of lymphoid (DAUDI, L-428, ATL-2, PB-1, 2A and Jurkat) cell cells after 3 h show chemotaxis; +, ,10% of cells after 3 h show lines. To avoid any influence of other, unidentified serum- chemotaxis; −, not detectable; ND, not done. derived factors, all experiments were performed under serum- free conditions. To our surprise, none of the chemokines

+ tested in this study inﬂuenced the proliferation of hemato- Table 6 Biological responses of CCR5 cell lines to stimulation by poietic cell lines (Figure 4). MIP-1β or RANTES

Cells Lineage Calcium flux Chemotaxis MIP-1β MIP-1β, RANTES Secretion of chemokines by human hematopoietic cell lines K-562 myeloid −− HEL myeloid −− DAMI myeloid −−Chemokines have recently been shown, by us and others, to DAUDI B cell −−be secreted by normal hematopoietic cells, and are thought BL-1 B cell ND − to play an important role in the cross-talk between cells.6 We 697 B cell −−hypothesized that hematopoietic cell lines can also secrete HS Sultan B cell −−HIV-related chemokines. We therefore collected media con- JB-6 T cell ND − − ditioned by selected cell lines and assayed for the presence Sez4 T cell ND β α β PB-1 T cell −−of the CCR5-related -chemokines (MIP-1 , MIP-1 and 2A T cell −−RANTES) using the sensitive ELISA assay. Since an ELISA assay Hs445 HD ND − is not commercially available for SDF-1, expression of SDF-1 was evaluated at the mRNA level by RT-PCR (Table 7). −, not detectable; ND, not done. We found that several hematopoietic cells lines, especially of B and T cell origin, secreted CCR5-related chemokines. MIP-1α was secreted by HS445, HUT102B, 15A, RS11846, observed that L428 and 15A cells that highly expressed PB-1, HsSultan and BC-1, whereas MIP-1β was produced by CXCR4 on the surface did not show calcium flux after stimu- HS445 and HUT102B, and RANTES was secreted by 15A, lation with SDF-1. Of note, all cell lines that displayed cal- HS445, RS11846 and HUT102B cells. In contrast, only three cium flux after stimulation by SDF-1 also displayed chemo- out of 19 cell lines evaluated in this study expressed mRNA for taxis to this chemokine (Table 6). In all, |40% of the CXCR4+ SDF-1 (Jurkat, HUT-102B and PB-1). The expression of SDF-1 cell lines did not respond to SDF-1 in the chemotaxis assay by HUT-102B cells was subsequently confirmed by Western (L428, BL-1, 15A, MOLT13 and Sez4). blot (data not shown).

Leukemia Hematopoietic cell lines and HIV M Majka et al 1826 Table 7 Endogenous secretion of chemokines (106 cells/1 ml/24 h) by human hematopoietic cell lines (pg/ml)

Cells Lineage MIP-1α MIP-1β RANTES SDF-1 (ELISA)a (ELISA)a (ELISA)a (mRNA)

K-562 myelo −−−− HL-60 myelo − 40 ± 23 −− UT-7 myelo 111 ± 26 −−− HEL erythr −−−− Dami mega −−−− BL-1 B cell 242 ± 54 −−− Nalm6 B cell − ND −− 697 B cell − ND −− HS Sultan B cell 740 ± 120 ND −− RS11846 B cell 2639 ± 432 ND 1437 ± 213 − 15A B cell 2504 ± 513 ND 3119 ± 237 − Jurkat T cell −−−+ Molt13 T cell −−−− PB-1 T cell 1986 ± 324 − 244 ± 34 + 2A T cell 63 ± 18 − 437 ± 78 − JB-6 T cell 60 ± 12 − 125 ± 35 − HUT102B T cell 2497 ± 616 1183 ± 320 1667 ± 251 + L428 HD 78 ± 22 − 1139 ± 320 − HS445 HD 2596 ± 509 1258 ± 347 2285 ± 860 −

The presence of CCR5 binding chemokines (MIP-1α, MIP-1β and RANTES) was evaluated by ELISA assay whereas expression of SDF- 1 mRNA was by RT-PCR. The data are presented as mean ± s.d. from four independent experiments. aThreshold of detection: MIP-1α .31 pg/ml; MIP-1β .31 pg/ml; RANTES .15 pg/ml. +, detectable; −, not detectable; ND, not done.

Figure 2 Calcium flux studies of Fura-2 loaded PB-1 (upper panel) and Jurkat (lower panel) cells. SDF-1α (500 ng/ml) (a, c) or MIP-1β (500 ng/ml) (b, d) were added and subsequently calcium flux was evaluated by spectrophotofluorimeter. Representative data of triplicate experiments are shown.

Endogenously secreted chemokines modulate the T lymphocytic cells (PB-1) (previously found to be permissive infectibility of cells by HIV strains to infection by both R5 and X4 HIV strains32) with the virus in the presence of conditioned media collected from either In view of the above observations we hypothesized that HEL (negative by ELISA for all β-chemokines tested and nega- endogenously secreted chemokines might somehow modulate tive by RT-PCR for SDF-1 mRNA), 15A (positive by ELISA for infection by R5 or X4 HIV. To test this hypothesis, we infected MIP-1α and RANTES, negative by RT-PCR for SDF-1 mRNA),

Leukemia Hematopoietic cell lines and HIV M Majka et al 1827

Figure 3 Chemotaxis studies of PB-1, ATL-2, 2A and Jurkat cells were conducted without (upper panel), or with 300 ng/ml of SDF-1β (middle panel) or 300 ng/ml of MIP-1β (lower panel). Responsiveness was determined by FACS. We found that all cells showed chemotaxis to SDF- 1α but not to MIP-1β. Representative data of triplicate experiments are shown.

or HUT102B cells (positive by ELISA for MIP-1α, MIP-1β and Discussion RANTES and positive by RT-PCR for SDF-1 mRNA). As a negative control the same conditioned media that were first heat To learn more about the role of HIV-related chemokine recep- inactivated were used. tors in human hematopoiesis we: (1) phenotyped various Infectivity and inhibition of infectivity were examined by human hematopoietic cell lines for expression of CXCR4 and both morphologic criteria (Figure 5) and by p24 antigen pro- CCR5; (2) studied biological responses of these receptors in duction (Figure 6). Conditioned media collected from 15A and calcium flux and chemotaxis assays; (3) investigated the effect HUT102B cells, which contain CCR5 binding chemokines, of chemokines on hematopoietic cell proliferation; (4) exam- significantly decreased the infection of PB-1 cells by the R5 ined whether chemokines are endogenously secreted by hem- HIV strain (Figure 5, upper panels c and e). Infection of these atopoietic cell lines; and finally (5) determined whether target cells, however, was not affected by conditioned endogenously secreted chemokines interfere with HIV medium from the β-chemokine-negative HEL cells (Figure 5, infection of hematopoietic cell lines. upper panel a). Conditioned media harvested from HUT102B We observed that several human hematopoietic cell lines cells which are positive for the CXCR4-binding SDF-1 mRNA express CD4 and HIV coreceptors CCR5 and CXCR4. We next inhibited infection of PB-1 cells by the X4 HIV strain (Figure identified several cell lines that should be susceptible to infec- 5, lower panel e). In contrast, conditioned media harvested tion by either X4 or R5 strains. Accordingly, C91PL, PB-1, from HEL and 15A cells (Figure 5, lower panels a and c), Sez4 and L540 cells, which are CD4+CXCR4+CCR5+, should which do not express SDF-1 mRNA, were non-protective. be permissive to X4 and R5 HIV. In contrast HL-60, Jurkat, Inactivation of the conditioned media by boiling resulted in MOLT4, ATL-2 and HUT102B cells, which are inhibition of this protective effect (Figure 5, upper panels d CD4+CXCR4+CCR5− and HEL, DAMI, UT-7, C10MJ2 and and f, lower panel f). Karpas 299 cells that are CD4+CXCR4−CCR5+, should be The results of p24 antigen production in these cultures is permissive to X4 or R5 HIV, respectively. shown in Figure 6. As above, PB-1 cells that were incubated Several human hematopoietic cell lines were previously with R5 HIV (BAL) (Figure 6, left panels) in the presence of shown to support HIV infection (HEL, Jurkat, DAMI and conditioned media from 15A (Figure 6b) or HUT102B (Figure MOLT4), and have been employed as experimental models in 6c) produced less p24 than cells incubated with heat-inacti- AIDS research.33–37 Here, we identified additional cell lines vated conditioned media from 15A and HUT102B lines (P , (PB-1, Sez-4, HUT102B, ATL-2, C91PL, L540, C10MJ2 and 0.0001). Similar analysis conducted using X4 HIV (HXB2) Karpas 299) which should also be useful tools for studying (Figure 6, right panels) showed an inhibiting p24 production various aspects of HIV biology. The possibility that human in cultures containing HUT102B (Figure 6c) conditioned hematopoietic cell lines could become infected by HIV media (P , 0.0001), but not that from HEL (Figure 6a) or 15A should always be kept in mind, especially in those labora- (Figure 6b). tories that have broad access to human blood products.

Leukemia Hematopoietic cell lines and HIV M Majka et al 1828

Figure 4 Selected human hematopoietic cells lines were maintained in artiﬁcial serum supplemented Iscove’s DMEM medium and then stimulated with various chemokines (SDF-1α, MIP-1α, MIP-1β or RANTES). Proliferation was determined by direct cell counts at days 4, 7 and 14. Results represent the mean of duplicate experiments.

Human hematopoietic cell lines maintained in these labora- MCP-1 on proliferation of 19 different myeloid leukemic cell tories could acquire a virus from cells from HIV-positive lines, but contradict a report from Sanchez et al,26 who found patients by cross-contamination and become a potential that SDF-1 strongly inhibits proliferation of murine hematopo- reservoir of HIV. ietic cells. This latter observation, however, is somehow unex- To our surprise, we noticed that the CD4 antigen, which is pected because SDF-1, which is a strong chemoattractant for a typical marker of T lymphocyte subsets is expressed on the human hematopietic progenitors, does not inhibit the prolifer- surface of several human myeloid cells lines (TF-1, HEL, HL- ation of human cells.21,23,24,39 Generally, the data on the effect 60, DAMI and UT-7). It has been reported that the CD4 anti- of chemokines on proliferation of both normal6,13,14,16,19,22,39 gen is present on normal human megakaryocytes,25 mono- and malignant6,18,26 hematopoietic cells are often contradic- cytes,38 and differentiating erythroblasts.7 The biological sig- tory and may depend on such things as: (1) the source of nificance of the CD4 antigen expression on both normal and chemokines; (2) modification of chemokine receptor function malignant human myeloid cells requires further studies. and their expression by factors released by hematopoietic Since the effects of chemokines on cell proliferation remain accessory cells; and (3) formation of inactive dimers by controversial, we incubated 10 different CXCR4+ and/or chemokines in solution.6,7,14,40 Proliferation assays in our stud- CCR5+ human hematopoietic lines with SDF-1, MIP-1α, MIP- ies were performed in a culture supplemented with chemi- 1β or RANTES. We did not detect any effect of these chemo- cally defined artificial serum to eliminate the influence of con- kines on proliferation either in serum-free or serum-sup- founding factors (usually present in different batches of animal plemented cultures, even if particular receptors were or human sera) and the cells were stimulated by commercially expressed and functional as assessed by calcium flux and/or available highly purified recombinant chemokines. chemotaxis assays. These data suggest that chemokines do not Furthermore, we determined here that the expression of a directly influence cell proliferation. particular chemokine receptor does not necessarily correlate Our data are in agreement with a recent report from Drexler with biological function. We observed that while several et al 8 who did not find any effect of RANTES, MIP-1α and human hematopoietic CXCR4+ cell lines responded to SDF-1

Leukemia Hematopoietic cell lines and HIV M Majka et al 1829

Figure 5 PB-1 cells infected with R5 HIV (BAL, upper panel) or X4 HIV (HXB2, lower panel) in the presence of conditioned medium collected from HEL (a and b), 15A (c and d) and HUT102B (e and f) cells. We employed freshly collected conditioned media (a, c, e) or conditioned media which were heat inactivated (b, d, f). The photographs were taken 10 days after infection. The experiment was repeated twice with similar results.

in chemotaxis and/or calcium ﬂux assays (eg HL-60, Jurkat, kines might protect normal human immunocompetent cells NALM6, ATL2, L428, PB1, 2A, BL-1, 697, SULTAN, 15A, (CD8+ and CD4+ T cells and NK cells) from HIV infection.42–47 MOLT13 and Sez4), none of the CCR5+ cell lines responded In agreement with these observations, we recently reported to stimulation by MIP-1α, MIP-1β or RANTES (eg K-562, HEL, that normal human hematopoietic cells (CD34+ cells and DAMI, Daudi, JB-6, BL-1, 697, SULTAN, Sez4, PB-1, 2A megakaryoblasts) also secrete HIV-related β-chemokines, and and Hs445). postulated that the latter may autoprotect hematopoietic cells The functional responses of malignant human hematopo- from infection by R5 HIV strains.48 However, the amount of ietic cells to stimulation by α- and β-chemokines are similar chemokine secreted by these cells is generally small. Luster to those previously reported by our laboratories using normal et al argued that since endogenously secreted β-chemokines human hematopoietic cells.39 Primary human hematopoietic are complexed with proteoglycans they may compete more cells (CD34+ cells or megakaryoblasts) displayed chemotaxis effectively with HIV binding to CCR5 than uncomplexed and calcium ﬂux after stimulation with SDF-1 but not with recombinant chemokines.43 MIP-1α or MIP-1β.39 The relative unresponsiveness of the In this study, we found that several human T and B cell CCR5 receptor on human hematopoietic cells could be poten- lines may secrete CCR5-binding β-chemokines as well as the tially explained by the differences in sulfation of the CCR5 CXCR4 ligand SDF-1α and that chemokines endogenously receptor,41 cross-desensitization of chemokine receptors,11 or secreted by established human hematopoietic cell lines, simi- by differences in the biological function of RGS proteins.10 larly to chemokines secreted by normal lympho–hematopo- Moreover, hematopoietic cells might also become resistant to ietic cells,42–47 might interfere with the infectability of the cells stimulation with exogenous chemokines if they endogenously by R5 and X4 HIV strains. Of note, as a target for HIV infection secrete chemokines.42–47 we employed PB-1 cells, which are infectable by both R5 and It has also been shown that endogenously secreted chemo- X4 HIV.32,48 Interestingly, however, as we found that these

Leukemia Hematopoietic cell lines and HIV M Majka et al 1830 ical point of view as we demonstrated that: (1) expression of chemokine receptors does not necessarily correlate with their functional responses in calcium flux and chemotaxis assays; (2) the CXCR4 receptor appears to be more biologically functional in human hematopoietic cells than the CCR5 receptor; and (3) chemokines do not influence proliferation of hematopoietic cell lines, even when the functionally relevant receptors are present on the cells. These data are also significant with respect to HIV research as: (1) we identified several human cell lines which may be employed as models for HIV infection studies; (2) we found that HIV-related receptors are expressed on many human hematopoietic cell lines (myeloid and lymphoid) and thus human hematopoietic cell lines may potentially serve as a reservoir of HIV; and (3) we demonstrate that the infectability of human hematopoietic cell lines may be automodulated by endogenously secreted chemokines. Finally, our data could be of special interest to all those researchers who are studying the role of chemokine– chemokine receptor axes in cell biology.

Acknowledgements

This paper was presented as an oral presentation at 41st Annual Meeting of American Society of Hematology, 1999 New Orleans, LA, December 3–7 and published in abstract form Blood 94: 618a, 1999 (abstr. suppl. 1). The authors are indebted to Dr Anna Janowska-Wieczorek from University of Alberta for critical comments. This work was supported by a NIH grant R01 HL61796–01 to MZR and R01 AI4083 to GNG.

References

1 Fauci AS. Host factors and the pathogenesis of HIV-induced disease. Nature 1996; 384: 529–534. 2 Hoxie JA. Hematologic manifestations of AIDS. In: Hoffman R, Benz EJ, Shatill SJ, Furie B, Cohen HJ, Silberstein LE (eds). Hema- tology Basic Principles and Practice. Churchill Livingstone: New York, 1995, pp 2171–2200. 3 Moses A, Nelson J, Bagby GC. The inﬂuence of human immuno- deﬁciency virus-1 on hematopoiesis. Blood 1998; 91: 1479–1495. 4 Lee B, Doranz BJ, Ratajczak MZ, Doms RW. An intricate web: chemokine receptors, HIV-1 and hematopoiesis. Stem Cells 1998; 16: 79–88. 5 Cairns JS, D’Souza MP. Chemokines and HIV-1 second receptors: the therapeutic connection. Nature Med 1998; 4: 563–568. 6 Rollins BJ. Chemokines. Blood 1997; 90: 909–928. 7 Lee B, Ratajczak J, Doms RW, Gewirtz AM, Ratajczak MZ. Coreceptor/chemokine receptor expression on human hematopoietic cells: biological implications for HIV-1 infection. Blood 1999; Figure 6 p24 antigen production in HIV infected cultures. p24 93: 1145–1156. HIV protein (pg/ml) was determined in conditioned media collected 8 Gutkind JS. The pathways connecting G protein-coupled receptors from PB-1 cells 10 days after infection with R5 HIV (BAL) or X4 HIV to the nucleus through divergent mitogen-activated protein kinase (HXB2) in the presence of fresh conditioned media (white bars) col- cascades. J Biol Chem 1998; 273: 1839–1842. lected from HEL (a), 15A (b) or HUT102B (c) cells or heat inactivated 9 Majka M, Ratajczak J, Kowalska MA, Ratajczak MZ. Binding of conditioned media collected from the same cells (black bars). Results stromal derived factor-1a (SDF-1a) to CXCR4 chemokine receptor represent the mean of duplicate experiments. *P , 0.0001 as in normal human megakaryoblasts but not in platelets induces compared to fresh medium. phosphorylation of mitogen-activated protein kinase p42/44 (MAPK), ELK-1 transcription factor and serine/threonine kinase AKT. Eur J Haematol 2000; 64: 164–172. 10 Dohlman HG, Thorner J. RGS proteins and signaling by hetero- trimeric G proteins. J Biol Chem 1997; 272: 3871–3874. cells also secrete some HIV-protective chemokines, we pre- 11 Lefkowitz PJ. G protein-coupled receptors. III. New roles for dicted that the relatively low secretion of RANTES and SDF- receptor kinases and b-arrestins in receptor signaling and desensitization. J Biol Chem 1998; 273: 18677–18680. 1 by these cells may explain why they remain permissive to 12 Lu Li, Xiao M, Grigsby S, Wang WX, Wu B, Shen RN, Broxmeyer infection by R5 and X4 HIV, respectively. HE. Comparative effects of suppressive cytokines on isolated sin- In conclusion, these data are important from the hematolog- gle CD34++ stem/progenitor cells from human bone marrow and

Leukemia Hematopoietic cell lines and HIV M Majka et al 1831 umbilical cord blood plated with and without serum. Exp Hematol Ratajczak MZ. An improved serum free system for cloning human 1993; 21: 1442–1446. ‘pure’ erythroid colonies. The role of the different growth factors 13 Broxmeyer HE, Sherry B, Cooper S, Lu L, Maze R, Beckmann MP, and cytokines on BFU-E formation by the bone marrow and cord Cerami A, Ralph P. Comparative analysis of the human macro- blood CD34+ cells. Folia Histochem Cytobiol 1998; 36: 55–60. phage inflammatory protein family of cytokines (chemokines) on 31 Ratajczak J, Zhang Q, Pertusini E, Wojczyk B, Wasik M, Ratajczak proliferation of human myeloid progenitor cells. J Immunol 1993; MZ. The role of insulin (INS) and insulin like growth factor-I (IGF- 150: 3448–3458. I) in regulating human erythropoiesis. Studies in vitro under serum 14 Mantel C, Kim YJ, Cooper S, Kwon B, Broxmeyer HE. Polymeriz- free conditions – comparison to other cytokines and growth fac- ation of murine macrophage inflammatory protein 1a inactivates tors. Leukemia 1998; 12: 371–381. its myelosuppressive effects in vitro: the active form is monomer. 32 Majka M, Lee B, Ratajczak J, Pertusini E, Honczarenko M, Kowal- Proc Natl Acad Sci USA 1993; 90: 2232–2236. ska MA, Wasik MA, Gewirtz AM, Ratajczak MZ. Expression and 15 Su S, Mukaida N, Wang J, Zhang Y, Takami A, Nakao S, Matsush- function of HIV-1 co-receptor on human hematopoietic cell lines. ima K. Inhibition of immature erythroid progenitor cell prolifer- Blood 1998; 92 (Suppl. 1): 166a. ation by macrophage inflammatory protein-1a by interacting 33 Yao XJ, Mouland AJ, Subbramanian RA, Forget J, Rougeau N, mainly with a C–C chemokine receptor, CCR1. Blood 1997; 90: Bergeron D, Cohen EA. Vpr stimulates viral expression and 605–611. induces cell killing in human immunodeficiency virus type-1 16 VanRanst PCF, Snoeck HW, Lardon F, Lenjou M, Nijs G, Weekx dividing Jurkat T cells. J Virol 1998; 72: 4686–4693. SFA, Rodrigus I, Berneman ZN, VanBockstaele DR. TGF-b and 34 Cane AM, Chang J, Naif H, Amato T, Cunningham AL, Chong BH. MIP-1a exert their main inhibitory activity on very primitive Human immunodeficiency virus-1 infection of megakaryocytic ++ − CD34 CD38 cells but show opposite effects on more mature cells. Annu Conf Australas Soc HIV Med 1996; 8: 39. + + CD34 CD38 human hematopoietic progenitors. Exp Hematol 35 Zauli G, Catani L, Gribellini D, Rec MC, Milani D, Borgatti P, 1996; 24: 1509–1515. Bassini A, La Placa M, Capitani S. The CD4 receptor plays essen- 17 Mayani H, Little MT, Dragowska W, Thornbury G, Lansdorp PM. tial but distinct roles in HIV-1 infection and induction of apoptosis Differential effects of the hematopoietic inhibitors MIP-1a, TGF- in primary bone marrow GPIIb/IIIa+ megakaryocytes and the HEL b, and TNF-a on cytokine-induced proliferation of subpopulations cell line. Br J Haematol 1995; 91: 290–298. + of CD34 cells purified from cord blood and fetal liver. Exp Hem- 36 DiFronzo NL, Pise-Masison CA, Fernandez-Larsson R, Holland atol 1995; 23: 422–427. CA. Viral determinants of HIV-1 sufficient to extend tropism to 18 Drexler HG, Zaborski M, Quentmeier H. Cytokine response pro- macrophages are distinct from the determinants that control the files of human myeloid factor-dependent leukemia cell lines. cytopathic phenotype in HL-60 cells. AIDS 1997; 11: 1681–1688. Leukemia 1997; 11: 701–708. 37 Ushijima H, Dairaku M, Honma H, Yamaguchi K, Shimizu H, 19 Patel VP, Kreider BL, Li Y, Li H, Leung K, Salcedo T, Nardelli B, Tsuchie H, Abe K, Yamamoto A, Hoshino H, Muller WE. Human Pippalla V, Gentz S, Thotakura R, Parmelee D, Gentz R, Gianni immunodeficiency virus infection in cells of myeloid-monocytic G. Molecular and functional chracterization of two novel human lineage. Microbiol Immunol 1991; 35: 487–492. C-C chemokines as inhibitors of two distinct classes of myeloid 38 Shen H, Cheng T, Preffer FI, Dombkowski D, Tomasson MH, progenitors. J Exp Med 1997; 185: 1163–1172. Golan DE, Yang O, Hofmann W, Sodroski JG, Luster AD, Scadden 20 Cook DN, Beck MA, Coffman TM, Kirby SL, Sheridan JF, Prangell AD. Intrinsic human immunodeficiency virus type 1 resistance of IB, Smithies O. Requirement of MIP-1a for an inflammatory hematopoietic stem cells despite coreceptor expression. J Virol response to viral infection. Science 1995; 269: 1583–1585. 1999; 73: 728–737. 21 Dutt P, Wang JF, Groopman JE. Stromal cell-derived factor-1a and 39 Majka M, Lee B, Honczarenko M, Ratajczak J, Douglas R, Kowal- stem cell factor/kit ligand share signaling pathways in hemopoietic ska MA, Silberstein L, Gewirtz AM, Ratajczak MZ. The stromal progenitors: a potential mechanism for cooperative induction of derived factor-1 (SDF-1) receptor – CXCR4 is expressed on the chemotaxis. J Immunol 1998; 161: 3652–3658. human erythroid progenitors and SDF-1 is an important chemotac- 22 Keller IR, Bartelmez SH, Sitnicka E, Ruscetti FW, Ortiz M, Gooya tic factor for early human erythroid cells. Comparison to other JM, Jacobsen SEW. Distinct and overlapping direct effects of chemokine receptor–chemokines axes. Stem Cells 2000; 18: macrophage inflammatory protein-1a and transforming growth 128–138. factor-b on hematopoietic progenitor/stem cell growth. Blood 40 Bonecchi R, Sozzani S, Stine JT, Luini W, D’Amico G, Allavena 1994; 84: 2175–2181. P, Chantry D, Mantovani A. Divergent effects of interleukin-4 and 23 Aiuti A, Webb IJ, Bleul C, Springer T, Gutierrez-Ramos JC. The interferon-g on macrophage-derived chemokine production: an chemokine SDF-1 is a chemoatractant for human CD34+ hemato- amplification circuit of polarized T helper 2 responses. Blood poietic progenitor cells and provides a new mechanism to explain 1998; 92: 2668–2671. the mobilization of CD34+ progenitors to peripheral blood. J Exp 41 Farzan M, Mirzabekov T, Kolchinsky P, Wyatt R, Cayabyab M, Med 1997; 185: 111–120. 24 Mohle R, Bautz F, Rafii S, Moore MA, Brugger W, Kanz L. The Gerard NP, Gerard C, Sodorski J, Choe H. Tyrosine sulfatation of chemokine receptor CXCR4 is expressed on CD34+ hematopoietic the amino terminus of CCR5 facilitates HIV-1 entry. Cell 1999; progenitors and leukemic cells and mediates transendothelial 96: 667–676. 42 Fehinger TA, Herbein G, Yu H, Para MI, Bernstein ZP, O’Brien migration induced by stromal cell-derived factor-1. Blood 1998; + 91: 4523–4530. WA, Caliguri MA. Natural killer cells from HIV-1 patients pro- 25 Kowalska MA, Ratajczak J, Hoxie J, Brass L, Gewirtz AM, Poncz duce C–C chemokines and inhibit HIV-1 infection. J Immunol M, Ratajczak MZ. Platelet and megakaryocytes express the HIV 1998; 161: 6433–6438. co-receptor CXCR4 on their surface but do not respond to stromal 43 Wagner L, Yang OO, Garcia-Zepeda EA, Ge Y, Kalmas SA, Walker β derived factor (SDF)-1. Br J Haematol 1999; 104: 220–229. BD, Pasternack MS, Luster AD. -chemokines are released from 26 Sanchez X, Cousins-Hodges B, Aguilar T, Gosselink P, Lu Z, Nav- HIV-1 specific cytolytic T-cell granules complexed to proteogly- arro J. Activation of HIV-1coreceptor (CXCR4) mediates myelo- cans. Nature 1998; 391: 908–911. suppression. J Biol Chem 1997; 272: 27529–27531. 44 Oliva A, Ninter AL, Vaccarezza M, Rubbert A, Catanzaro A, Moir 27 Nagasawa T, Hirota S, Tachibana K, Takakura N, Nishikawa S, S, Monaco J, Ehler L, Mizelli S, Jackson R, Li Y, Romano JW, Fauci Kitamura Y, Yoshida N, Kikutani H, Kishimoto T. Defects of B-cell AS. Natural killer cells from human immunodeficiency virus lymphopoiesis and bone-marrow myelopoiesis in mice lacking the (HIV)-infected individuals are an important source of CC-chemo- CXC chemokine PBSF/SDF-1. Nature 1996; 382: 635–638. kines and suppress HIV-1 entry and replication in vitro. J Clin 28 Zou Y, Kottmann AH, Kuroda M, Taniuchi I, Littman DR. Function Invest 1998; 102: 223–231. of the chemokine receptor CXCR4 in haematopoiesis and in cer- 45 Fehniger TA, Herbein G, Yu H, Para MI, Bernstein ZP, O’Brien ebellar development. Nature 1998; 393: 595–599. WA, Caliguri MA. Natural killer cells from HIV-1+ patients pro- 29 Drexler HG, Matsuo Y. Guidelines for the characterization and duce CC-chemokines and inhibit HIV-1 infection. Blood 1998; 92 publication of human malignant hematopoietic cell lines. (Suppl. 1): 544a. Leukemia 1999; 13: 835–842. 46 Cocchi F, DeVico AL, Garzino-Demo A, Arya SK, Gallo RC, Luso 30 Ratajczak J, Marlicz W, Machalinski B, Pertusini E, Czajka R, P. Identification of RANTES, MIP-1α, and MIP-1β as the major

Leukemia Hematopoietic cell lines and HIV M Majka et al

1832 + HIV suppressive factors produced by CD8 T cells. Science 1995; 48 Majka M, Rozmyslowicz T, Lee B, Pietrzkowski Z, Gaulton GN, 270: 1811–1815. Silberstein L, Ratajczak MZ. Bone marrow CD34+ cells and 47 Saha K, Bentsman G, Chess L, Volsky DJ. Endogenous production megakaryoblasts secrete β-chemokines; implications for infect- of b-chemokines by CD4+, but not CD8+, T-cell clones correlates ability by M-tropic human immunodeﬁciency virus (R5 HIV). J with the clinical state of human immunodeﬁciency virus type-1 Clin Invest 1999; 104: 1739–1749. (HIV-1)-infected individuals and may be responsible for blocking infection with non-syncytium-inducing HIV-1 in vitro. J Virol 1998; 72: 876–881.

Leukemia