Molecular and Functional Characterization of  Two Novel Human C-C Chemokines as Inhibitors of  Two Distinct Classes of Myeloid Progenitors

HBSS, IMDM, RPMI 1640, and MEM tissue culture media, l-glutamine, penstrep, and poly (A) RNA purification kits were purchased from GIBCO BRL (Gaithersburg, MD). Heat-inactivated fetal bovine serum (FBS), histopaque 1077 and 1119, DMSO, platelet-activating factor (PAF), and PMA were from Sigma Chemical Co. (St. Louis, MO). MyeloCult™ H5100 and MethoCult™ H4535 media were purchased from Stem Cell Technologies Inc. (Vancouver, Canada). Bacto-agar was from Difco Labs. (Detroit, MI). EXCELL 400 medium was from J.R. Scientific (Woodland, CA). Tissue culture grade plasticware was from Costar Corp. (Cambridge, MA). QBEND/10 CD34 cell isolation kit, RS and CS columns, and VarioMac were purchased from Miltenyi Biotech Inc. (Sunnyvale, CA). The cDNA ZAP Express synthesis kits were from Stratagene Corp. (La Jolla, CA). Poly (A) RNA from various tissues was purchased from Clontech (Palo Alto, CA). Nylon membranes for Northern blot analysis were from Zetabind (Cuno, Inc., Meriden, CT). Restriction enzymes and protease inhibitors were from Boehringer Mannheim (Indianapolis, IN). POROS CM20 and HS50 columns were from PerSeptive Biosystems (Framingham, MA). Sephacryl S-200 was from Pharmacia BioProcess Technology ASB (Uppsala, Sweden). ProBlott™ membranes and Blot Cartridges^TM were from Perkin-Elmer Corp./ Applied Biosystems, (Foster City, CA). PA-1 column was from Dionex Corp. (Sunnyvale, CA). Limulus amebocyte lysate assay was from Associates of Cape Cod, Inc. (Woods Hole, MA). The intracellular fluorescence probes were from Molecular Probes Inc. (Eugene, OR).

The mAbs were purchased from Becton Dickinson (San Jose, CA). The mAb HIT3a anti–human CD3 was from PharMingen (San Diego, CA). Human and mouse recombinant cytokines were purchased from R&D Sys., Inc. (Minneapolis, MN). MCP-4 was produced as reported (16).

The cell lines Sf9 (CRL-1711), K562 (CCL-243), HL60 (CRL-1964), THP-1 (TIB-202), Jurkat (TIB152), Molt-3 (CRL-1552), RS4-11 (CRL-1873), CCRF-CEM (CCL-119), KG1 (CCL-246), and KG1a (CCL-246.1) were from American Type Culture Collection (Rockville, MD). Most of the cell lines were cultured in RPMI 1640 supplemented with 10% FBS, 2 mM l-glutamine, and penstrep. KG1 and KG1a cells were grown in IMDM with 10% FBS. For the chemical induction studies, ∼2 × 10⁶ K562 or HL-60 cells were distributed in T-75 flasks with 20 ml of fresh media with or without 1% DMSO or 10 μM PMA, and processed for total RNA at various time points. Sf9 cells were grown at 27°C in EX-CELL 400 medium with 1% FBS. The Autographa californica virus was from PharMingen.

Poly (A) RNA was purified from a variety of tissues or cell lines using purification kits according to the manufacturer's recommendations. Total RNA was isolated using the guanidine isothiocynate–CsCl₂ method (17). All the cDNA libraries were then constructed by using the ZAP-cDNA kit according to the manufacturer's instructions. Poly (A) RNA (1 μg) from tissues or total RNA from cell lines (10 μg) were fractionated on a 1% agaroseformaldehyde gel, transferred to nylon membranes, and hybridized under high stringency conditions with ³²P-cDNA probes as previously described (18).

Bacterial colonies containing plasmids with inserts were selected, lysed, and the cDNA insert was amplified by PCR. Expressed sequence tag (EST) clones were discovered as reported (19, 20). The EST clones used in this study were the result of a collaboration between The Institute for Genomic Research (Gaithersburg, MD) and Human Genome Sciences, Inc. (Rockville, MD). For sequence analysis and assembly, Sequencher Software (version 2.1.1; Gene Codes Corp., Ann Arbor, MI) was used.

The coding sequences of MPIF-1 and MPIF-2 were amplified by PCR during which unique restriction sites (BglII and Asp 718) were introduced and then cloned into the baculovirus transfer vector A2. This vector contains the β-galactosidase gene of Escherichia coli to facilitate the identification of recombinant viruses that form blue plaques. The Sf9 cells were infected at a multiplicity of infection of two and the supernatant of the infected culture was collected after 72 h. The recombinant MPIF-1 and MPIF-2 were purified in the presence of protease inhibitors (20 μg/ml pefabloc SC, 1 μg/ ml leupeptin, 1μg/ml E-64, and 1 mM EDTA) on a POROS HS 50 cation exchange column. The column was eluted with 500 mM NaCl in 40 mM Na acetate buffer at pH 5.5; the chemokine positive fractions were pooled and diluted with a 40 mM Na acetate buffer, pH 5.5, to adjust NaCl concentration to 100 mM. This mixture was applied to a POROS CM20 cation exchange column and eluted with a linear salt gradient (0.1–1 M NaCl). The positive fractions were pooled and applied to a Sephacryl S-200 column. The resulting preparations were >95% pure as determined by SDS-PAGE. Endotoxin was monitored by the limulus amebocyte lysate assay.

The NH₂-terminal amino acid sequence of purified MPIF-1 and MPIF-2 was analyzed by an ABI-494 sequencer (Perkin-Elmer). The glycosylation of each chemokine was determined by a Dionex carbohydrate analyzer (21).

Peripheral blood mononuclear neutrophils, eosinophils, T lymphocytes, monocytes, and PBMC were purified from single donor leukopacks (American Red Cross, Rockville, MD) as previously described (22–24).

Resting T lymphocytes were enriched by negative selection by passing PBMC on T cell columns (R&D Sys.) that retained monocytes and B lymphocytes. Resting CD4⁺ and CD8⁺ T lymphocytes were 90% pure. Enriched T lymphocytes (CD4⁺ and CD8⁺) were activated by 16 h incubation in RPMI medium in the presence of 10 U/ml IL-2 in 100-mm dishes coated with HIT3a mAb anti–CD3 (1 μg/ml). Elutriated monocytes were 70–80% pure as assessed by CD14 positivity. Eosinophils were >95% pure as assessed on stained cytospun preparations. CD34⁺ cells were isolated from 45 ml of cord blood collected with 5 mM EDTA and 20 U/ml heparin. After 10 min centrifugation at 200 g the buffycoat was resuspended in PBS supplemented with 5 mM EDTA (PBS/EDTA) and centrifuged on histopaque 1077. The cells were collected, washed, resuspended in cold PBS/EDTA with 0.5% BSA, and subjected to the QBEND/10 kit and magnetic cell sorting according to the manufacturer's instructions. This procedure yielded a 95% CD34⁺ population of cells.

Bone marrow cells were isolated from 6–8-wk-old C57Bl/6 female mice (Jackson Labs., Bar Harbor, ME) according to Metcalf (25). In brief, cells were obtained from the femur and tibia by flushing with IMDM 5% FBS and centrifuged at 750 g for 30 min over a cushion of histopaque 1119. Cells sedimenting at the interface were washed with IMDM medium, plated in 10-cm dishes, and incubated for 1 h at 37°C to remove plastic adherent cells. The nonadherent cell population was used in clonogenic assays.

Cells were washed in HBSS with 0.1% BSA (HBSS/BSA) and resuspended in the same medium at 2 × 10⁶/ml with 1 μM calcein-AM. After 30 min at 37°C, the cells were washed in HBSS/BSA, resuspended in the same medium at 8 × 10⁶/ml, and 25 μl of this suspension was dispensed into each upper chamber of a 96-well chemotaxis plate filter (Neuro Probe, Cabin John, MD). Different concentrations of the chemotactic agents were distributed in the bottom chamber of each well. Cells were allowed to migrate for 45–90 min through the polycarbonate filter between upper and bottom chambers (5–8 μm pores; polyvinylpyrrolidone-free), and the number of migrated cells (both attached to the bottom surface of the filter as well as in the bottom chamber) were quantitated using a fluorescence plate reader (Cytofluor PerSeptive Biosystems). The ratio between the number of cells migrated in the presence of chemokines and the number of cells migrated in the presence of buffer control is defined as the chemotactic index.

Monocytes were loaded for 30 min at 37°C with 4 μM Indo-1/acetoxymethylester per 5 × 10⁶ cells/ml in prewarmed buffer (1 mM CaCl₂, 2 mM MgSO₄, 5 mM glucose, 10 mM Hepes buffer, pH 7.4). Cells were then washed, resuspended in the same buffer at 5 × 10⁵ cells/ml, and stimulated with the indicated chemokines at 37°C. Eosinophils were loaded with 2.5 μM fura-2/5 × 10⁶ cells/ml in prewarmed buffer (1 mM CaCl₂, 1 mM MgCl₂, 125 mM NaCl, 5 mM KCl, 0.5 mM glucose, 20 mM Hepes, pH 7.4, 0.025% BSA). Cells were washed in loading buffer and resuspended at 10⁶ cells/ml. The fluorescence signal induced by the changes in the intracellular Ca²⁺ was measured on a fluorescence spectrophotometer (F-2000; Hitachi Instruments Inc., San Jose, CA) by monitoring Indo-1 excitation at 330 nm, emission at 405 and 485 nm, and fura-2 excitation at 340 and 380 nm, emission at 510 nm.

High proliferative potential colony-forming cells (HPP-CFC) and low proliferative potential CFC (LPP-CFC) colony-formation assays were performed in a two-layered agar culture system as described by Bradley and Hodgson (26). In brief, the bottom layer was prepared in 3.5-cm diam dishes with 1 ml of MEM supplemented with 20% FBS, 0.5% Difco agar, 7.5 ng/ml mouse (m)IL-3, 150 ng/ml mouse stem cell factor (mSCF), 7.5 ng/ml human monocyte (hM)-CSF, and 15 ng/ml mIL-1α. Chemokines were incorporated into the bottom agar where indicated. This layer was then overlayed with 0.5 ml of murine bone marrow cell suspension to have 1,500 cells/dish in MEM with 20% FBS and 0.3% agar. The dishes were then incubated for 14 d in a tissue culture incubator (37°C, 88% N₂, 5% CO₂, and 7% O₂) and colonies were scored under an inverted microscope. In these conditions HPP-CFC and LPP-CFC gives rise to colonies that are >5 mm and <1 mm diam, respectively (27, 28). In some experiments, impact of MPIF-1 and MPIF-2 on the growth of burst-forming unit–erythroid (BFU-E), CFU-E, and CFU–granulocyte and monocyte (GM) was also determined. In brief, 5 × 10⁴ bone marrow cells were suspended in 1 ml of MethoCult™ semisolid medium and plated in 3.5-cm dishes. For CFU-E assay, the medium was supplemented with 3 U/ml h erythropoietin (Epo) and the dishes were incubated for 2 d. The hemoglobinized colonies (4–16 cells/ colony) were identified by acid benzidine stain and scored under an inverted microscope. For BFU-E and CFU-GM assay, the medium was supplemented with 5 ng/ml mIL-3, 10 ng/ml mGMCSF, and 3 U/ml hEpo. The dishes were incubated for 11 d and the colonies were scored under an inverted microscope.

CD34⁺ cord blood cells (5 × 10⁴ cells/ml) were cultured in myelocult^TM medium supplemented with 10 ng/ml hIL-3 and 50 ng/ml hSCF for 4 d at 37°C in a tissue culture incubator (5% CO₂, 7% O₂, and 88% N₂). The resulting populations of progenitors were suspended (10³ cells/ml) in Methocult^TM semisolid medium supplemented with a cocktail of cytokines (10 ng/ml hIL-3, 10 ng/ml hGM-CSF, 50 ng/ml hSCF, 3U/ml hEpo, and 10 ng/ml h thrombopoietin), and plated in the presence or absence of chemokines. Colonies were allowed to grow by incubating the dishes at 37°C in a humidified incubator (5% CO₂, 7% O₂, and 88% N₂) for 14 d, and colonies were scored under an inverted microscope.

The ESTs representing MPIF-1 and MPIF-2 cDNAs were identified in the Human Genome Sciences, Inc. data base on the basis of the C-C motif and the homology with known β-chemokines. The MPIF-1 cDNA clone was isolated from a human aortic endothelial library. Sequencing of both strands of the insert revealed a full coding region for a 120–amino acid protein, of which the 21 NH₂-terminal amino acids represent the signal peptide (Fig. 1,A). In comparison with the known human β-chemokines, MPIF-1 shows six cysteine residues and is most closely related to hMIP-1α, exhibiting 51% identity and 67% similarity over a stretch of 99 amino acids (Fig. 2). The amino acid sequence of MPIF-1 revealed no N-linked glycosylation sites. The MPIF-2 cDNA clone was isolated from an activated monocyte library and encodes for a 119–amino acid protein, of which the 26 NH₂-terminal amino acids represent the signal peptide (Fig. 1 B). MPIF-2 is a β-chemokine with four cysteine residues that exhibits 39% identity and 59% similarity to MCP-3, and 42% identity and 65% similarity to MIP-1α over a stretch of 93 amino acids. The overall comparison of MPIF-1 with MPIF-2 revealed 30% identity and 50% similarity to one another (not shown).

MPIF-1 ESTs were most frequently found in the liver, lung, and bone marrow cDNA libraries, but MPIF-2 ESTs were restricted to the activated monocyte and T lymphocyte cDNA libraries. To confirm this observation, Northern blot analysis was performed. As shown in Fig. 3,A, MPIF-1 cDNA hybridized to a 0.8-kb mRNA species that is most abundant in the adult lung and liver (Fig. 3,A, lanes 7 and 8), and is present at moderate levels in fetal liver (Fig. 3,A, lane 4), adult bone marrow and placenta (Fig. 3,A, lanes 11 and 13). Analysis of various cell lines showed that MPIF-1 mRNA is readily detectable in myelomonocytic cell lines (Fig. 3 B). Similar analysis with MPIF-2 cDNA turned out negative. The MPIF-2 mRNA was only detectable by reverse transcriptase-PCR in monocytes activated with GM-CSF for 96 h and in T lymphocytes activated by mAb anti–CD3 (HIT3a) in the presence of IL-2 (not shown).

Recombinant MPIF-1 and MPIF-2 proteins were expressed in Sf9 insect cells infected with a recombinant baculovirus. SDS-PAGE analysis of a purified MPIF-1 preparation revealed a single Coomassie blue–stained band with a molecular mass of 11–12 kD (Fig. 4,A). Edman degradation analysis showed that the MPIF-1 protein was processed to yield a protein with RVTKDAE as the NH₂-terminus and an expected molecular mass of ∼11.2 kD, thus confirming that the observed sequence is identical to that predicted from the cDNA sequence (Fig. 4 C). Consistent with the lack of N-glycosylation sites no mannose or N-acetylglucosamine were detected in the MPIF-1 protein.

Amino acid sequence analysis indicated that the NH₂ terminus of the MPIF-2 protein is not blocked and begins with VVIPSP (Fig. 4,C). This predicts the calculated molecular mass of the mature MPIF-2 protein to be 10.5 kD, which is in agreement with the SDS-PAGE analysis of Coomassie blue–stained protein (Fig. 4 B). MPIF-2 contains up to 0.9 moles of N-acetylglucosamine or mannose per mole of the protein. No galactose or N-acetylgalactosamine were detected, suggesting the absence of O-linked oligosaccharides.

The chemotactic properties of MPIF-1 and MPIF-2 were studied on purified resting and anti-CD3–activated T lymphocytes, neutrophils, and monocytes. The activity on T lymphocytes and monocytes was tested in comparison with MCP-1, which is a potent attractant for these cells. The activity on neutrophils was tested, and compared with IL-8. Several experiments were performed using cells obtained from different donors, and representative experiments are summarized in Fig. 5.

On resting T lymphocytes, MPIF-1 reached its maximum activity at 10 ng/ml, and was more potent and three times more effective than MCP-1. In contrast, MPIF-2 and MCP-1 showed comparable potency and efficacy (Fig. 5,C). Both MPIF-1 and MPIF-2 were negative on anti-CD3– activated T lymphocytes (Fig. 5,D). MPIF-1 induced migration of monocytes at concentrations ranging from 100 ng/ml to 1 μg/ml, but this activity was not observed with all the monocyte preparations tested. MPIF-2 was negative on monocytes (Fig. 5,B). Both MPIF-1 and MPIF-2 induced some migration of neutrophils. Although minimal, this activity was consistently evident at 50 ng/ml, whereas IL-8 was already four times more potent at 10 ng/ml (Fig. 5 A).

The capacity of MPIF-1 and MPIF-2 to induce a rapid and transient rise of cytosolic Ca²⁺ concentration [Ca²⁺]_i was assessed on monocytes, and eosinophils. MPIF-1 was active on monocytes, and MPIF-2 produced changes of [Ca²⁺]_i on eosinophils. On monocytes, a significant rise of the [Ca²⁺]_i was observed using 10 ng/ml MIP-1α, and the maximum value was recorded at 100 ng/ml MIP-1α. In similar conditions, 1 μg/ml MPIF-1 induced only 50% of the change of [Ca²⁺]_i that was observed with 100 ng/ml MIP-1α. In the desensitization experiments, the [Ca²⁺]_i changes after repeated stimulation were monitored in monocytes to assess receptor usage by MPIF-1. The stimulation of monocytes with 100 ng/ml MIP-1α abolished responsiveness to 1 μg/ ml MPIF-1, but not to 100 ng/ml MCP-4 (Fig. 6,B). When monocytes were first exposed to 1 μg/ml MPIF-1 and then to 100 ng/ml MIP-1α followed by 100 ng/ml MCP-4, MPIF-1 was able to abolish MIP-1α response without affecting the ability of the same cells to respond to MCP-4 (Fig. 6 A).

On eosinophils, the magnitude of the [Ca²⁺]_i rise increased with the concentration of chemokines tested. A maximum response was recorded using 500 ng/ml eotaxin or 1 μg/ml MPIF-2 (data not shown). At the maximal concentration of 1 μg/ml, the response to MPIF-2 was ∼85% of the eotaxin response, while MCP-1 at 1 μg/ml was negative. In desensitization experiments, the stimulation of eosinophils with 1 μg/ml eotaxin or MCP-4 abolished responsiveness to 1 μg/ml MPIF-2, but not to 10⁻⁶ M PAF (Fig. 7, A and D). When eosinophils were first exposed to 1 μg/ml MPIF-2 and then to 1 μg/ml eotaxin or MCP-4, MPIF-2 was found to be able to abolish eotaxin or MCP-4 response without affecting the ability of the same cells to respond to PAF (Fig. 7, B and C).

The effects MPIF-1 and MPIF-2 on hematopoietic cells was studied on mouse low density bone marrow cells plated in methylcellulose- or agar-containing growth medium supplemented with appropriate combinations of cytokines. No colonies were detected either in the absence of cytokines or in the presence of either MPIF-1 or MPIF-2 alone. A concentration of 50 ng/ml MPIF-1 reduced the frequency of CFU-GM and LPP-CFC colonies to 30% of that found in the control cultures (Table 1). This inhibitory effect of MPIF-1 appeared to be specific, as MPIF-1 had no effect on colony formation by CFU-E, BFU-E, and HPP-CFC (Table 1). In contrast, 50 ng/ml MPIF-2 produced 80% reduction in the numbers of HPP-CFC colonies compared to those found in the control cultures without effects on colony formation by CFU-E, BFU-E, CFU-GM, and LPP-CFC (Table 1). In the same conditions, both MCP-4 (50 ng/ml) and MIP-1α (100 ng/ml) had no effect on colony formation (Table 1 and Fig. 8). The inhibitory effect of MPIF-1 and MPIF-2 was dose dependent (Fig. 9). MPIF-2 reduced HPP-CFC colonies to 50% of the control at 10 ng/ml (Fig. 9,A), while MPIF-1 reduced LPP-CFC colonies to 50% of the control at 20 ng/ml (Fig. 9 B). This experiment confirmed the specificity of MPIF-1 and MPIF-2 inhibitory effect on LPP-CFC and HPP-CFC, respectively.

To determine the impact of MPIF-1 and MPIF-2 on human hematopoietic cells, CD34⁺ cells were isolated from cord blood and cultured for 4 d in the presence of IL-3 and SCF to generate hematopoietic progenitors. A limiting dilution of this cell population was then cultured in semisolid medium with multiple cytokines with or without test chemokines. Data from two representative experiment are shown in Table 2. Colony formation by CFU-GM and CFU-Mix in the presence of either 50 ng/ml MPIF-1 or 50 ng/ml MPIF-2 was reduced by 30–40% compared to those found in the control cultures (Table 2). Neither 50 ng/ml of MIP-1α nor MCP-4 affected CFU-GM and CFU-Mix (Table 2).

In a large-scale DNA sequencing effort, two novel human β-chemokines were discovered and named Ckβ-8 or MPIF-1 and Ckβ-6 or MPIF-2 for their activity on human and mouse myeloid progenitors. The cDNA encoding MPIF-1 has been cloned from a human aortic endothelial library, and the message for this chemokine is mainly expressed by macrophages and is detectable in lung, liver, bone marrow, placenta, and fetal liver. MPIF-1 is not NH₂ terminally blocked, and shares 51% identity and 67% similarity with MIP-1α. The chemotactic activity of MPIF-1 is very potent on resting T lymphocytes, significant on monocytes, and weak but consistent on neutrophils. Finally, MPIF-1 has a notable in vitro inhibitory effect on LPP-CFC, the committed progenitors that give rise to granulocyte and monocyte lineages and reduces CFU-GM, but has no significant effects on HPP-CFC multipotent hematopoietic progenitors. The experiments performed with human cord blood CD34⁺-derived cells show that MPIF-1 inhibits both CFU-Mix and CFU-GM, and confirm that our novel β-chemokine acts on similar targets of both murine and human origins. MPIF-1 does not show inhibitory activity on BFU-E or megakaryocyte precursors. Although the desensitization experiments suggest that MPIF-1 and MIP-1α may share the same receptor on monocytes, they are functionally unrelated. MPIF-1 and MIP-1α are active on different functional stages of T lymphocytes and hematopoietic progenitors where they may use distinct receptors. Finally, in comparison with MIP-1α, MPIF-1 is a more potent inhibitor of hematopoietic progenitors. In our study, MPIF-1 at 10–50 ng/ml produces 50–80% inhibition of mLPP-CFC and 50% inhibition of hCFU-GM and hCFU-Mix (CFU-GEMM), whereas MIP-1α at 100 ng/ ml is ineffective. These data are in agreement with previously published results showing inhibitory effect of MIP1α for BFU-E, CFU-GM, CFU-GEMM, and CFU-Megs at concentrations ranging from 100 to 500 ng/ml (15, 29, 30). Other studies show that a single addition of MIP-1α up to 1 μg/ml to the liquid or semisolid system has no effect on GM-CFU, BFU-E, and multilineage precurors. Only the daily addition of 100 ng/ml MIP-1α for 6 d produces inhibition (31, 32).

The recently cloned MRP-2 is a mouse β-chemokine that shares some characteristics with MIP-1α and shows interesting similarities to MPIF-1. It suppresses in vitro colony formation by mouse and human BFU-E, CFU-GM, and CFU-GEMM progenitors (33). MPIF-1 and MRP-2 are 42% identical at the amino acid level and show six cysteine residues that are conserved with respect to their location within both proteins. Due to the fact that the MRP-2 and MPIF-1 cDNAs exhibit only 67% identity and MPIF-1 does not inhibit BFU-E, it is unclear if MPIF-1 represents the human counterpart of MRP-2. The structural and functional properties we described for MPIF-1 suggest that this novel β-chemokine may be grouped with MIP-1α, MIP-1β, RANTES, and mMRP-2.

The novel β-chemokine MPIF-2 is weakly expressed by activated T lymphocytes and GM-CSF–treated macrophages, does not have the NH₂ terminus blocked, and exhibits 40% identity and 59% similarity to MCP-3, and 42% identity and 65% similarity to MIP-1α. MPIF-2 attracts resting T lymphocytes, activates eosinophils, has minimal activity on neutrophils, and is negative on monocytes and activated T lymphocytes. On eosinophils, MPIF-2, eotaxin, and MCP-4 induce comparable transient rise of [Ca²⁺]_i. Finally, MPIF-2 has a notable in vitro inhibitory effect on mHPP-CFC multipotent hematopoietic precursors without affecting the LPP-CFC–committed progenitors that give rise to granulocyte and monocyte lineages. Although the desensitization experiments suggest that MPIF-2, eotaxin, and MCP-4 may share the same receptor on eosinophils, MPIF-2 may use a different receptor on multipotent hematopoietic precursors. In fact, MCP-4 does not affect HPP-CFC multipotent hematopoietic precursors. The structural and functional properties we described for MPIF-2 show that this novel β-chemokine shares characteristics with both the MCP and MIP chemokines.

Presently, it is difficult to assess the physiological relevance of the novel β-chemokines we described in this paper. During early stages of inflammation, wound healing, and delayed-type hypersensitivity reactions, locally activated macrophages, granulocytes, lymphocytes, and endothelial cells release β-chemokines (i.e., MCP-1, MCP-3, MIP-1α, MIP-1β, and RANTES), cytokines (i.e., IL-1, IL-3, IL-5, IL-6), and growth factors (i.e., G-CSF, GM-CSF) to recruit mature leukocytes and to stimulate the bone marrow production of new leukocytes (34–36). It is tempting to speculate that MPIF-1 and MPIF-2 may play an important role promoting and then limiting these processes. Once released by activated T lymphocytes, monocytes, and tissue macrophages, MPIF-1 and MPIF-2 could induce the migration of these same cell types to the local microenvironment. More intriguing, the potential of MPIF-1 and MPIF-2 to reduce the capacity of the hematopoietic progenitor pool to produce mature elements could limit the local accumulation of proinflammatory cells. The inhibitory activity of MPIF-1 and MPIF-2 suggests that these novel chemokines may have the potential to protect the hematopoietic progenitors from the cytotoxic effects of the antiblastic drugs used in cancer therapy.

We wish to thank Vina Patil-Koota, Radhika Uppaluri, and Jeffrey Carrell for their excellent technical assistance. We would also like to thank the Shady Grove Hospital (Rockville, MD) for their provision of human cord blood. The contributions of the Human Genome Sciences, Inc. and The Institute of Genomic Research (Rockville, MD). Sequencing and Bioinformatics departments are appreciated. The human tissues in this study were supplied by the Cooperative Human Tissue Network (Fredrick, MD), which is funded by the National Cancer Institute.

Note added in proof. The cDNA sequence for MPIF-2 shown in Fig. 1 B has a typographical error at nucleotide 182, which is printed c rather than g. This correction results in a change of the amino acid at position 61 from a glycine (G) to an alanine (A). In addition, the MPIF-2 cDNA sequence predicts a phenylalanine at position 73. However, data from mass spectrophotometry and sequencing studies indicates that there is a serine at position 73 instead of a phenylalanine in the protein preparation used in this study.