Antiangiogenesis–based cancer therapies, specifically those targeting the VEGF-A/VEGFR2 pathway, have been approved for subsets of solid tumors. However, these therapies result in an increase in hematologic adverse events. We surmised that both the bone marrow vasculature and VEGF receptor–positive hematopoietic cells could be impacted by VEGF pathway–targeted therapies. We used a mouse model of spontaneous breast cancer to decipher the mechanism by which VEGF pathway inhibition alters hematopoiesis. Tumor-bearing animals, while exhibiting increased angiogenesis at the primary tumor site, showed signs of shrinkage in the sinusoidal bone marrow vasculature accompanied by an increase in the hematopoietic stem cell–containing Lin-cKit+Sca1+ (LKS) progenitor population. Therapeutic intervention by targeting VEGF-A, VEGFR2, and VEGFR3 inhibited tumor growth, consistent with observed alterations in the primary tumor vascular bed. These treatments also displayed systemic effects, including reversal of the tumor-induced shrinkage of sinusoidal vessels and altered population balance of hematopoietic stem cells in the bone marrow, manifested by the restoration of sinusoidal vessel morphology and hematopoietic homeostasis. These data indicate that tumor cells exert an aberrant systemic effect on the bone marrow microenvironment and VEGF-A/VEGFR targeting restores bone marrow function. Cancer Res; 76(3); 517–24. ©2015 AACR.

Cancer has local (juxtacrine/paracrine) and distal (paracrine/endocrine) effects. The study of its local effects, particularly upon the proliferative and poorly functional vasculature, has led to its description as a wound that does not heal (1). Although cancer can have significant pathologic systemic effects, such as anemia and cachexia, little is known about its influence upon distal vasculature. One particularly important non-tumor vascular bed is the sinusoidal vasculature of the bone marrow, due to its importance in hematopoiesis (2). Hematopoietic stem cells (HSC) in the bone marrow exist primarily in a quiescent state, but can be compelled to enter the bloodstream by agents that interfere with the interactions between HSCs and their microenvironments (3). HSCs are a subset of the hematopoietic stem/progenitor cell (HSPC) bone marrow compartment, which includes multiple subsets with varying differentiative capabilities. HSPC-derived cells can be recruited to distant organs where they promote metastasis, as well as into primary tumors (4). Systemic conditions such as diabetes and estradiol treatment can target the subset of bone marrow HSPCs capable of reconstituting the hematopoietic system (5, 6), and recent work shows that tumor burden increases bone marrow HSPC numbers (7), raising the possibility that an important source of HSPC-derived cells recruited by tumors may be the niches of the bone marrow.

In general, HSPC-derived hematopoietic cells enter body tissues via VEGFR2-expressing blood vessels and exit via VEGFR2- and VEGFR3-expressing lymphatic vessels. The sinusoidal vasculature is unique in that it mediates traffic in both directions, serving as a gatekeeper to virtually all fluid and cellular elements that enter and exit the bone marrow. In contrast to many other vascular beds, it lacks a thick basement membrane, is virtually unsupported by pericytes, and expresses VEGFR2, VEGFR3, and VEGFR1 (8, 9). We wondered whether the sinusoidal vessels, which serve as a niche for HSPCs, would be affected by increased circulating levels of VEGF-A (as seen in cancer) or anti-VEGF pathway inhibitors, a classification that includes many cancer drugs both approved and in development due to the important role angiogenesis plays in the tumor microenvironment. In addition to their well-documented effects on local tumor vasculature, antiangiogenic therapies could also affect the bone marrow sinusoidal bed in its role as the vascular niche, potentially affecting hematopoietic homeostasis and contributing to the increase in adverse hematopoietic events with these therapies (10, 11). In addition, previous studies have shown that some hematopoietic cells express VEGF receptors, and of special interest to the biology of HSPC-derived cells we considered the impact of anti-VEGFR1 on VEGFR1-expressing LKS HSPCs.

Mice

Tumor volume of female MMTV-PyT mice was calculated as 0.52*(smaller dimension)2*(larger dimension). Blood was analyzed on a Hemavet (Drew Scientific). All animal studies were approved by the Beth Israel Deaconess Medical Center Institutional Animal Care and Use Committee.

Immunohistochemistry

Tumors were fixed in 4% paraformaldehyde and paraffin embedded. One femur per animal was decalcified in 10% EDTA after fixation. Four-micron sections were baked for 1 hour at 60°F. Antigens were retrieved on deparaffinized, rehydrated slides using proteinase K or Universal Decloaker (Biocare Medical). Slides were washed with 0.05% Tween/TBS and blocked with Protein Block (Dako). Tumor and femur sections were incubated with biotinylated antibodies to VEGFR3 (AFL4; eBioscience), CD31 (Becton-Dickinson), or LYVE-1 (Abcam). Sections were subsequently washed, blocked with hydrogen peroxide, incubated with streptavidin-HRP, and exposed with DAB (Dako). After additional washes, the sections were counterstained with hematoxylin (Dako), dehydrated, and cleared with xylene prior to mounting with cytoseal XYL (ThermoScientific). Entire tissue sections were imaged at 20× with the Aperio ScanScopeXT. Regions of interest were defined to avoid quantification within regions with folded tissue or high background staining and to restrict analysis to bone marrow regions of the femur. Endothelial area was defined as stained cells only. Average endothelial area per vessel was quantified using custom-built algorithms in Definiens image analysis software. Results are aggregated per animal and expressed as mean per group ± SEM. Statistical differences between experimental and control groups were measured by the two-tailed Student t test.

Antibody treatment

Female MMTV-PyT mice with total tumor volumes of 0.4 to 0.6 cm3 were injected i.p. with 800 μg of control rat IgG (Jackson ImmunoResearch), anti-VEGFR1 (MF1), anti-VEGFR2 (DC101), anti-VEGFR3 (mF4-31C1) or 400 μg of anti-VEGF-A (G6-31; ImClone Systems) three times a week for 1 week or until total tumor volume reached approximately 1 cm3.

Flow cytometry

Femurs were flushed with cold PBS, treated with lysis buffer (Quality Biological), and counted by a hemacytometer. Cells were blocked with anti-Fc (eBioscience) then incubated with directly conjugated antibodies against lineage markers (B220, CD4, CD8, GR1, TER119, and CD11b), Sca-1, and cKit (eBioscience; Biolegend). Antibodies against CD34, Flt3, CD48, and CD150 further characterized the population. For cell cycle, LKS-stained cells were permeabilized with 0.03% saponin and stained with 7-AAD and pyronin Y. Cells were acquired on a FACSCanto II and analyzed using FACSDiva (BD Biosciences).

Competitive bone marrow transplants

Female FVB recipients were given 9 Gy of gamma irradiation. After 24 hours, the animals were reconstituted by tail-vein injection with 5 × 105 cells from a GFP+ donor and varying numbers of cells from a GFP tumor-bearing MMTV-PyT or tumor-negative littermate control. After 12 weeks of reconstitution, leukocytes were isolated from the blood and stained with CD45 (eBioscience). A positive competitive result was defined as a percentage of CD45+ GFP+ cells below the 95% confidence level of control animals that received GFP+ bone marrow exclusively. Data were analyzed using a Poisson statistic, and the number of competitive repopulating units was estimated using LCalc (StemCell Technologies).

Quantitative real-time PCR

RNA was isolated from bone marrow and lungs using TRIzol (Invitrogen) and cDNA synthesis performed with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). RNA was quantified using Sybr Green (Applied Biosystems) on a 7500 Fast Real-Time PCR System (Applied Biosystems) using primers against PyT (F5′-AGCCCGATGACAGCATATCC-3′, R5′-GGTCTTGGTCGCTTTCTGGA-3′; Integrated DNA Technologies). Values were normalized against 18S.

ELISA

Blood was collected in EDTA-containing tubes (BD Biosciences) and spun for 20 minutes at 2,000 rpm at 4°C. Concentrations of SDF-1/CXCL12, SCF, G-CSF, VEGF-A, VEGF-C, and VEGF-D (R&D Systems) in plasma and/or serum were analyzed according to the manufacturer's instructions. GM-CSF, IL6, and TNFα were assessed with the Multi-Analyte ELISArray Kit (SABiosciences).

To examine the impact of cancer and antiangiogenic therapy on the bone marrow sinusoidal vessels and HSPC population, we used the MMTV-PyT model of triple-negative breast cancer. The spontaneous tumors arising in this model are metastatic and display comparable morphology to human disease (12). Because some but not all tumors contain VEGFR3+ blood vessels (13), we first profiled the expression of VEGFR2 and VEGFR3 in our model. VEGFR3 was widely expressed throughout primary MMTV-PyT tumors, primarily on blood vessels that were identified by lack of LYVE-1 positivity and presence of luminal erythrocytes (Fig. 1A). We examined tumor growth in response to selective neutralizing antibodies to all three VEGF receptors. As expected, anti-VEGFR2 and anti-VEGF-A antibodies resulted in a significant (and comparable) delay in tumor growth, while administration of anti-VEGFR1 antibodies had no such effect. This latter result is consistent with previous reports showing failure of VEGFR1 inhibition to block tumor angiogenesis and growth (14). Surprisingly, anti-VEGFR3 also resulted in delayed tumor growth (Fig. 1B) without inducing necrosis like anti-VEGFR2 treatment (Fig. 1C). Antibody treatment did not reduce the density of CD31+ vessels overall, although nonspecific staining of necrotic tissue may have camouflaged anti-VEGFR2 effects. However, the inhibitory effects of anti-VEGFR2 and anti-VEGFR3 antibodies on tumor progression were associated with significant shrinkage of endothelial area per vessel (Fig. 1C and D). These findings suggest that anti-VEGFR3 antibodies inhibit tumor growth in this model, at least in part, by attenuating vascular endothelial cell signaling in tumor blood vessels (15, 16).

Figure 1.

Morphologic and growth effects of anti-VEGFR2/3 treatment on MMTV-PyT primary tumor. A, VEGFR3+ vessels in the primary tumor were identified as members of the blood vasculature through similarity to CD31+ vessels, lack of LYVE-1 positivity, and presence of circulating blood cells (×20; scale bar, 100 μm). B, animals bearing tumors with total volume of 0.4 to 0.6 cm3 received antibody injections 3 times per week; animals were sacrificed when total tumor volume reached approximately 1.0 cm3. Inhibition of VEGF-A, VEGFR2, or VEGFR3 significantly delayed primary tumor growth, while VEGFR1 inhibition had no effect (n = 4–12). C and D, treatment with anti-VEGFR2/3 did not change CD31+ vessel density, but resulted in decreases in per-vessel endothelial area (×20; n = 3–5).

Figure 1.

Morphologic and growth effects of anti-VEGFR2/3 treatment on MMTV-PyT primary tumor. A, VEGFR3+ vessels in the primary tumor were identified as members of the blood vasculature through similarity to CD31+ vessels, lack of LYVE-1 positivity, and presence of circulating blood cells (×20; scale bar, 100 μm). B, animals bearing tumors with total volume of 0.4 to 0.6 cm3 received antibody injections 3 times per week; animals were sacrificed when total tumor volume reached approximately 1.0 cm3. Inhibition of VEGF-A, VEGFR2, or VEGFR3 significantly delayed primary tumor growth, while VEGFR1 inhibition had no effect (n = 4–12). C and D, treatment with anti-VEGFR2/3 did not change CD31+ vessel density, but resulted in decreases in per-vessel endothelial area (×20; n = 3–5).

Close modal

We next asked whether the tumor per se and/or its treatment with anti-VEGFR2 and anti-VEGFR3 antibodies influences the morphology of the bone marrow sinusoids. Compared with healthy controls, the sinusoid vessels of tumor-bearing animals appeared subtly smaller and more uniform (Fig. 2A) with a significant decrease in per-vessel endothelial area (Fig. 2B). Treatment with anti-VEGFR3 decreased the number of vessels, indicating an increased dependence on VEGFR3 signaling in the bone marrow compared with the primary tumor (Fig. 2B). However, treatment with anti-VEGFR2 or anti-VEGFR3 antibodies was able to partially rescue per-vessel endothelial area of the remaining vessels, suggesting that aberrant VEGFR family signaling can alter bone marrow vasculature without spurring angiogenesis (Fig. 2A and B).

Figure 2.

Morphologic effect of tumor burden and anti-VEGFR2/3 treatment on bone marrow vasculature. A, bone marrow from control and MMTV tumor–bearing animals treated with IgG, anti-VEGFR2, or anti-VEGFR3 was stained against VEGFR3 to identify the sinusoid vasculature (×20). B, quantification showed significant decreases in VEGFR3+ vessel density with VEGFR3 inhibition and partial rescue of per-vessel endothelial area with anti-VEGFR2/3 treatment (n = 3–5).

Figure 2.

Morphologic effect of tumor burden and anti-VEGFR2/3 treatment on bone marrow vasculature. A, bone marrow from control and MMTV tumor–bearing animals treated with IgG, anti-VEGFR2, or anti-VEGFR3 was stained against VEGFR3 to identify the sinusoid vasculature (×20). B, quantification showed significant decreases in VEGFR3+ vessel density with VEGFR3 inhibition and partial rescue of per-vessel endothelial area with anti-VEGFR2/3 treatment (n = 3–5).

Close modal

Recent work has demonstrated that the majority of HSCs are located in close proximity to the sinusoidal vasculature (2), which regulates HSC engraftment and hematopoiesis (9, 17). HSC subtypes can be identified by flow cytometry as part of the LineagecKit+Sca1+(LKS) HSPC compartment. Bone marrow from tumor-bearing MMTV-PyT mice displayed an increase in the percentage (data not shown) and absolute number (Fig. 3A) of LKS HSPCs, consistent with recent work in the MMTV-neuOTI/OTII model (7). In contrast to the study by Sio and colleagues (7), there was no evidence of increased extramedullary hematopoiesis in this model, with constant numbers of nucleated cells and LKS HSPCs in the spleen at all tumor volumes (data not shown). The increase in bone marrow LKS HSPCs was present even at small total tumor volumes and was independent of total bone marrow cellularity (Fig. 3A). This expanded LKS population in the bone marrow of tumor-bearing mice was hyperproliferative (Fig. 3B) and more differentiated, as shown by a relative decrease of less differentiated, CD34Flt3LKS long-term HSCs (LT-HSC) and an increase of more differentiated, CD34Flt3+LKS short-term HSCs (ST-HSCs; Fig. 3C). This shift toward a more differentiated LKS population was also observed using the markers CD48CD150+ (data not shown). Competitive bone marrow transplants revealed a decrease in functional LT-HSCs, with tumor-bearing animals displaying a significant decrease of over twofold in competitive repopulating units (CRU; Fig. 3D).

Figure 3.

The LKS progenitor population is dysregulated in the bone marrow of tumor-bearing MMTV-PyT mice. A, bone marrow isolated from tumor-bearing MMTV-PyT animals at various total tumor volumes was analyzed by flow cytometry. Tumor-bearing animals at total tumor volumes as small as 0.2 to 0.4 cm3 showed an increased number (*, P < 0.05; **, P = 0.005) of LKS cells in the bone marrow. Numbers of cells were calculated by multiplying the percentage of cells by the total number of cells in the bone marrow, as counted by hemacytometer (n = 4–9). B, staining with 7-AAD and pyronin Y revealed a relative decrease in G0–G1 phases and a relative increase in S–G2–M phases within the LKS population (n = 3; *P < 0.05). C, LKS cells were further characterized by staining against CD34 and Flt3. Tumor-bearing mice showed relative decreases in the less-differentiated CD34Flt3 LKS population and a relative increase in the more-differentiated CD34+Flt3 LKS population (n = 5–6; *, P < 0.04). D, lethally irradiated animals underwent competitive reconstitution with bone marrow from control or tumor-bearing mice versus GFP+ bone marrow. After 12 weeks of reconstitution, animals were sacrificed, and the leukocyte fraction was isolated from blood and stained for CD45. Increased competition from the unlabeled animals correlated with a decreased fraction of GFP+ cells. A positive competitive result is defined as a CD45+GFP+ fraction below the lower bound of the 95% confidence level of animals that were reconstituted with only GFP+ bone marrow. Data were analyzed using the Poisson statistic (L-Calc; *, P < 0.03).

Figure 3.

The LKS progenitor population is dysregulated in the bone marrow of tumor-bearing MMTV-PyT mice. A, bone marrow isolated from tumor-bearing MMTV-PyT animals at various total tumor volumes was analyzed by flow cytometry. Tumor-bearing animals at total tumor volumes as small as 0.2 to 0.4 cm3 showed an increased number (*, P < 0.05; **, P = 0.005) of LKS cells in the bone marrow. Numbers of cells were calculated by multiplying the percentage of cells by the total number of cells in the bone marrow, as counted by hemacytometer (n = 4–9). B, staining with 7-AAD and pyronin Y revealed a relative decrease in G0–G1 phases and a relative increase in S–G2–M phases within the LKS population (n = 3; *P < 0.05). C, LKS cells were further characterized by staining against CD34 and Flt3. Tumor-bearing mice showed relative decreases in the less-differentiated CD34Flt3 LKS population and a relative increase in the more-differentiated CD34+Flt3 LKS population (n = 5–6; *, P < 0.04). D, lethally irradiated animals underwent competitive reconstitution with bone marrow from control or tumor-bearing mice versus GFP+ bone marrow. After 12 weeks of reconstitution, animals were sacrificed, and the leukocyte fraction was isolated from blood and stained for CD45. Increased competition from the unlabeled animals correlated with a decreased fraction of GFP+ cells. A positive competitive result is defined as a CD45+GFP+ fraction below the lower bound of the 95% confidence level of animals that were reconstituted with only GFP+ bone marrow. Data were analyzed using the Poisson statistic (L-Calc; *, P < 0.03).

Close modal

To determine if the hyperproliferative, pro-differentiated phenotype of LKS HSPCs could be mediated locally by metastasized tumor cells in the bone marrow, quantitative real-time PCR was performed on lung and bone marrow from tumor-bearing mice. This assay identified metastasized tumor cells expressing PyT mRNA in lung tissue even from mice bearing tumors of only 0.2 to 0.4 cm3 (Fig. 4A), a pattern similar to the increased LKS population in the bone marrow. However, no significant burden of metastasized tumor cells was found in the bone marrow at any tumor volume (Fig. 4A). Therefore, we entertained the likely possibility that the effect of tumor on sinusoidal vasculature and HSPCs in the bone marrow was mediated by soluble factors. Leading candidates that we considered included VEGF-A, which promotes cycling and mobilization of hematopoietic progenitor cells from the bone marrow (18, 19), and G-CSF, which is used therapeutically to mobilize HSCs from the bone marrow (3). Indeed, plasma levels of both VEGF-A and G-CSF were significantly increased in tumor-bearing mice (Fig. 4B). The increased levels of G-CSF greatly affected the downstream differentiation of HSPCs in this model, with tumor-bearing mice displaying an 8-fold increase in circulating neutrophil concentration (data not shown). CXCL12/SDF-1 was another likely candidate given its role in HSC retention and mobilization (3), but the levels of CXCL12 in plasma (Fig. 4B) and serum (data not shown) were unchanged at even maximum total tumor volumes. Circulating levels of SCF, GM-CSF, IL6, and TNFα also remained unchanged (data not shown).

Figure 4.

Tumor effect on hematopoietic progenitor cells in the bone marrow is mediated by VEGF-A and can be rescued with anti-VEGF-A or anti-VEGFR3 treatment. A, lungs and bone marrow of mice bearing tumors of various sizes and their littermate controls were assayed for PyT mRNA by quantitative real-time PCR. B, plasma levels of VEGF-A, G-CSF, and CXCL12 were assayed by ELISA (n = 3–7). Animals with a total tumor burden of 0.4 to 0.6 cm3 were injected with antibodies against VEGF-A (G6), VEGFR1 (MF1), VEGFR2 (DC101), VEGFR3 (mF4-31C1). Anti-VEGF-A, VEGFR2, or VEGFR3 treatment significantly reduced the percentage of LKS cells in the S–M–G2 stages of the cell cycle. Evaluation of LKS or control IgG subsets by flow cytometry revealed two patterns of HSPC population response. C, anti-VEGF-A and anti-VEGFR3 treatments reduced CD34Flt3 LKS LT-HSCs and LSKs overall, while anti-VEGFR1 and anti-VEGFR2 treatment increased CD34+Flt3 LKS ST-HSCs and CD34+Flt3+ MPP populations (n = 6–16). D, model of tumor and VEGF family inhibitor actions in the bone marrow in vivo.

Figure 4.

Tumor effect on hematopoietic progenitor cells in the bone marrow is mediated by VEGF-A and can be rescued with anti-VEGF-A or anti-VEGFR3 treatment. A, lungs and bone marrow of mice bearing tumors of various sizes and their littermate controls were assayed for PyT mRNA by quantitative real-time PCR. B, plasma levels of VEGF-A, G-CSF, and CXCL12 were assayed by ELISA (n = 3–7). Animals with a total tumor burden of 0.4 to 0.6 cm3 were injected with antibodies against VEGF-A (G6), VEGFR1 (MF1), VEGFR2 (DC101), VEGFR3 (mF4-31C1). Anti-VEGF-A, VEGFR2, or VEGFR3 treatment significantly reduced the percentage of LKS cells in the S–M–G2 stages of the cell cycle. Evaluation of LKS or control IgG subsets by flow cytometry revealed two patterns of HSPC population response. C, anti-VEGF-A and anti-VEGFR3 treatments reduced CD34Flt3 LKS LT-HSCs and LSKs overall, while anti-VEGFR1 and anti-VEGFR2 treatment increased CD34+Flt3 LKS ST-HSCs and CD34+Flt3+ MPP populations (n = 6–16). D, model of tumor and VEGF family inhibitor actions in the bone marrow in vivo.

Close modal

We then tested the ability of anti-VEGF pathway inhibitors to reverse the hyperproliferative, pro-differentiated phenotype of HSPCs in the bone marrow. Although VEGFR1 inhibition did not affect primary tumor growth, VEGFR1 was included because it is expressed on LKS HSPCs as well as bone marrow endothelium (9, 20). We treated MMTV-PyT animals bearing tumors of 0.4 to 0.6 cm3 total tumor volume with antibodies against VEGF-A, VEGFR1, VEGFR2, VEGFR3, or control IgG three times per week until the total tumor volume reached 1.0 cm3. Anti-VEGF-A treatment reversed the tumor-induced expansion of the LKS HSPC compartment, evidenced by normalization of the cell cycle (Fig. 4C). Reduction in LT-HSC numbers confirmed that anti-VEGF-A treatment halted the inappropriate proliferation of that subset. Anti-VEGF-A–treated bone marrow, like tumor stimulated bone marrow, contained a high number of more differentiated, phenotypically marked ST-HSCs but a normal number of multipotent progenitors (MPP), a downstream subset that retains multilineage differentiation potential (Fig. 4C), suggesting that differentiation from ST-HSC to MPP in this context is slow and not controlled by VEGF-A. In contrast, VEGFR1 inhibition did not restore normal HPSC differentiation, even showing additional increases in ST-HSCs and MPPs (Fig. 4C).

The ability of VEGF-A and the inability of VEGFR1 inhibition to rescue the hyperproliferative, pro-differentiated LKS phenotype suggested that the sinusoidal vasculature may mediate communication between distant tumor and bone marrow HSPCs. We have shown that transgenic hyperactivation of Akt in endothelial cells results in an increase in bone marrow CRUs (21), and vasculature also mediates the establishment and regression of extramedullary hematopoietic sites in a VEGF-overexpressing tumor model (22). In support of this hypothesis, treatment with anti-VEGFR2, which is expressed on endothelium but not on LKS cells, restored LKS cell-cycle balance. However, like anti-VEGFR1 treatment, VEGFR2 blockade resulted in an increase in both ST-HSCs and MPPs (Fig. 4C). Treatment with anti-VEGFR3, which is also expressed on sinusoid endothelium but not on LKS cells, restored cell-cycle balance without further buildup of ST-HSCs or MPPs, and like VEGF-A blockade reduced LT-HSC and overall LKS numbers to control levels (Fig. 4C). In summary, we observed most complete reversal of tumor-induced progenitor buildup with anti-VEGF-A and anti-VEGFR3, consistent with decreased tumor growth, suggesting that these treatments counteract tumor-induced proliferation and differentiation of LT-HSCs (Fig. 4D). Anti-VEGFR2 treatment resulted in decreased HSPC proliferation and restoration of sinusoid structure, but also increased ST-HSC and MPP populations. Similar increases in ST-HSC and MPP populations with anti-VEGFR1 treatment suggest that expression of VEGFR1 on HSPCs and/or VEGFR1 and VEGFR2 on bone marrow vasculature may be required for terminal differentiation or exit of differentiated cells from the bone marrow, causing a buildup of these more differentiated cells in the bone marrow (Fig. 4D).

We confirm recent work showing that primary tumors are able to communicate long-distance with HSPCs in the bone marrow (7). Although the effect of other tumor-derived factors cannot be ruled out, this communication appears to be mediated by the effects of circulating VEGF-A on the sinusoidal vasculature of the bone marrow, as targeting of vascular-specific VEGF receptors can reverse the aberrant entrance of HSPCs into proliferative and differentiative phases of the cell cycle. In addition to the anti-VEGF-A (e.g., bevacizumab) and VEGFR non-selective TKIs (e.g., sunitinib, sorafenib, axitinib) approved for cancer therapy, there are multiple anticancer therapies under investigation for VEGFR1, VEGFR2, and VEGFR3 signaling. A selective fully human antibody to VEGFR2 (ramucirumab) has recently shown benefits in overall survival in gastric (10) and lung cancers, and selective antibodies to VEGFR3 (IMC-3C5) and VEGFR1 (icrucumab) are in early-phase trials for multiple solid tumors (www.clinicaltrials.gov). A better understanding of the role of the sinusoid vasculature may lead to increased antitumor treatment efficacy, specific targeting of hematopoietic malignancies that reside and proliferate in the bone marrow, improvements in HSC recruitment techniques for stem cell transplantation, and, given the importance of bone marrow–derived cells in preparing the metastatic niche (4), therapies that target the formation of metastases.

No potential conflicts of interest were disclosed.

Conception and design: R.K. O'Donnell, S. Rafii, W.C. Aird, L.E. Benjamin

Development of methodology: R.K. O'Donnell, B. Falcon, C. Perruzzi

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R.K. O'Donnell, B. Falcon, C. Perruzzi

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R.K. O'Donnell, B. Falcon, J. Hanson, W.E. Goldstein

Writing, review, and/or revision of the manuscript: R.K. O'Donnell, B. Falcon, J. Hanson, C. Perruzzi, W.C. Aird, L.E. Benjamin

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C. Perruzzi

Study supervision: W.C. Aird, L.E. Benjamin

This work was supported by the NIH (grant CA131152) and Eli Lilly and Company.

1.
Dvorak
HF
. 
Tumors: Wounds that do not heal
.
N Engl J Med
1986
;
315
:
1650
9
.
2.
Ding
L
,
Saunders
TL
,
Enikolopov
G
,
Morrison
SJ
. 
Endothelial and perivascular cells maintain haematopoietic stem cells
.
Nature
2012
;
481
:
457
63
.
3.
Nervi
B
,
Link
DC
,
DiPersio
JF
. 
Cytokines and hematopoietic stem cell mobilization
.
J Cell Biochem
2006
;
99
:
690
705
.
4.
Wels
J
,
Kaplan
RN
,
Rafii
S
,
Lyden
D
. 
Migratory neighbors and distant invaders: tumor-associated niche cells
.
Genes Dev
2008
;
22
:
559
74
.
5.
Illing
A
,
Liu
P
,
Ostermay
S
,
Schilling
A
,
de Haan
G
,
Krust
A
, et al
Estradiol increases hematopoietic stem and progenitor cells independent of its actions on bone
.
Haematologica
2012
;
97
:
1131
5
.
6.
Orlandi
A
,
Chavakis
E
,
Seeger
F
,
Tjwa
M
,
Zeiher
AM
,
Dimmeler
S
. 
Long-term diabetes impairs repopulation of hematopoietic progenitor cells and dysregulates the cytokine expression in the bone marrow microenvironment in mice
.
Basic Res Cardiol
2010
;
105
:
703
12
.
7.
Sio
A
,
Chehal
MK
,
Tsai
K
,
Fan
X
,
Roberts
ME
,
Nelson
BH
, et al
Dysregulated hematopoiesis caused by mammary cancer is associated with epigenetic changes and Hox gene expression in hematopoietic cells
.
Cancer Res
2013
;
73
:
5892
904
.
8.
Kopp
H-G
,
Avecilla
S
,
Hooper
AT
,
Rafii
S
. 
The bone marrow vascular niche: home of HSC differentiation and mobilization
.
Physiology
2005
;
20
:
349
56
.
9.
Hooper
AT
,
Butler
JM
,
Nolan
DJ
,
Kranz
A
,
Iida
K
,
Kobayashi
M
, et al
Engraftment and reconstitution of hematopoiesis is dependent on VEGFR2-mediated regeneration of sinusoidal endothelial cells
.
Cell Stem Cell
2009
;
4
:
263
74
.
10.
Wilke
H
,
Muro
K
,
Van Cutsem
E
,
Oh
SC
,
Bodoky
G
,
Shimada
Y
, et al
Ramucirumab plus paclitaxel versus placebo plus paclitaxel in patients with previously treated advanced gastric or gastro-oesophagael junction adenocarcinoma (RAINBOW): a double-blind, randomised phase 3 trial
.
Lancet Oncolo
2014
;
15
:
1224
35
.
11.
von Minckwitz
G
,
Puglisi
F
,
Cortes
J
,
Vrdoljak
E
,
Marschner
N
,
Zielinski
C
, et al
Bevacizumab plus chemotherapy versus chemotherapy alone as second-line treatment for patients with HER2-negative locally recurrent or metastatic breast cancer after first-line treatment with bevacizumab plus chemotherapy (TANIA): an open-label, randomised phase 3 trial
.
Lancet Oncol
2014
;
15
:
1269
78
.
12.
Lin
EY
,
Jones
JG
,
Li
P
,
Zhu
L
,
Whitney
KD
,
Muller
WJ
, et al
Progression to malignancy in the polyoma middle T oncoprotein mouse breast cancer model provides a reliable model for human diseases
.
Am J Pathol
2003
;
163
:
2113
26
.
13.
Smith
NR
,
Baker
D
,
James
NH
,
Ratcliffe
K
,
Jenkins
M
,
Ashton
SE
, et al
Vascular endothelial growth factor receptors VEGFR-2 and VEGFR-3 are localized primarily to the vasculature in human primary solid cancers
.
Clin Cancer Res
2010
;
16
:
3548
61
.
14.
Kushner
EJ
,
Bautch
VL
. 
Building blood vessels in development and disease
.
Curr Opin Hematol
2013
;
20
:
231
6
.
15.
Lakkonen
P
,
Waltari
M
,
Holopainen
T
,
Takahashi
T
,
Pytowski
B
,
Steiner
P
, et al
Vascular endothelial growth factor receptor 3 is involved in tumor angiogenesis and growth
.
Cancer Res
2007
;
67
:
593
9
.
16.
Petrova
TV
,
Bono
P
,
Holnthoner
W
,
Chesnes
J
,
Pytowski
B
,
Sihto
H
, et al
VEGFR-3 expression is restricted to blood and lymphatic vessels in solid tumors
.
Cancer Cell
2008
;
13
:
554
6
.
17.
Yao
L
,
Yokota
T
,
Xia
L
,
Kincade
PW
,
McEver
RP
. 
Bone marrow dysfunction in mice lacking the cytokine receptor gp130 in endothelial cells
.
Blood
2005
;
106
:
4093
5101
.
18.
Pitchford
SC
,
Furze
RC
,
Jones
CP
,
Wengner
AM
,
Rankin
SM
. 
Differential mobilization of subsets of progenitor cells from the bone marrow
.
Cell Stem Cell
2009
;
4
:
62
72
.
19.
Tashiro
K
,
Nonaka
A
,
Hirata
N
,
Yamaguchi
T
,
Mizuguchi
H
,
Kawabata
K
. 
Plasma elevation of vascular endothelial growth factor leads to the reduction of mouse hematopoietic and mesenchymal stem/progenitor cells in the bone marrow
.
Stem Cells Dev
2014
;
18
:
2202
10
.
20.
Gerber
H-P
,
Malik
AK
,
Solar
GP
,
Sherman
D
,
Liang
XH
,
Meng
G
, et al
VEGF regulates haematopoietic stem cell survival by an internal autocrine loop mechanism
.
Nature
2002
;
417
:
954
8
.
21.
Kobayashi
H
,
Butler
JM
,
O'Donnell
R
,
Kobayashi
M
,
Ding
BS
,
Bonner
B
, et al
Angiocrine factors from Akt-activated endothelial cells balance self-renewal and differentiation of haematopoietic stem cells
.
Nat Cell Biol
2010
;
12
:
1046
56
.
22.
Xue
Y
,
Chen
F
,
Zhang
D
,
Lim
S
,
Cao
Y
. 
Tumor-derived VEGF modulates hematopoiesis
.
J Angiogenesis Res
2009
;
I
:
9
.