Human Adipose-Tissue Derived Stromal Cells in Combination with Hypoxia Effectively Support Ex Vivo Expansion of Cord Blood Haematopoietic Progenitors

The optimisation of haematopoietic stem and progenitor cell expansion is on demand in modern cell therapy. In this work, haematopoietic stem/progenitor cells (HSPCs) have been selected from unmanipulated cord blood mononuclear cells (cbMNCs) due to adhesion to human adipose-tissue derived stromal cells (ASCs) under standard (20%) and tissue-related (5%) oxygen. ASCs efficiently maintained viability and supported further HSPC expansion at 20% and 5% O2. During co-culture with ASCs, a new floating population of differently committed HSPCs (HSPCs-1) grew. This suspension was enriched with СD34+ cells up to 6 (20% O2) and 8 (5% O2) times. Functional analysis of HSPCs-1 revealed cobble-stone area forming cells (CAFCs) and lineage-restricted colony-forming cells (CFCs). The number of CFCs was 1.6 times higher at tissue-related O2, than in standard cultivation (20% O2). This increase was related to a rise in the number of multipotent precursors - BFU-E, CFU-GEMM and CFU-GM. These changes were at least partly ensured by the increased concentration of MCP-1 and IL-8 at 5% O2. In summary, our data demonstrated that human ASCs enables the selection of functionally active HSPCs from unfractionated cbMNCs, the further expansion of which without exogenous cytokines provides enrichment with CD34+ cells. ASCs efficiently support the viability and proliferation of cord blood haematopoietic progenitors of different commitment at standard and tissue-related O2 levels at the expense of direct and paracrine cell-to-cell interactions.


Introduction
Cord blood haematopoietic stem and progenitor cells (cbHSPCs) have attracted considerable interest as a full value alternative to bone marrow HSPCs. The number of cbHSPC transplants of hypoxia, improved expansion of CD34 + HSPCs in the presence of osteoblasts [24] and bone marrow [1,25,26] MSCs was demonstrated.
Thus, the data on ex vivo HSPC expansion in the simulated haematopoietic microenvironment at tissue-related low O 2 and on bone marrow MSCs as a stromal layer provide a reason to conclude that these conditions may considerably modulate HSPC properties.
At present, most studies have implemented purified immuno-selected CD34 + HSPCs for ex vivo expansion. This may lead to artificial depletion of the heterogeneous HSPC population, as this antigen is expressed on only some subsets of HSPCs [27]. To enrich selection of the HSPC population from cord blood, we applied a functional approach at the expense of HSPC adhesion to the ASC layer and further expansion of the adhered HSPCs and their progeny [28]. In this study, we evaluated for the first time the potency of ASCs to support viability and selfrenewal/commitment of expanded HSPCs (progeny of cells which were adhered to unmanipulated cbMNCs) at standard (20%) and tissue-related (5%-hypoxia) O 2 .

Preparation of Cells
Adipose tissue samples were obtained in the frame of Scientific Agreement from multidisciplinary clinic "Souz" (Moscow, Russia) after elective liposuction procedures under local anaesthesia from healthy patients with written informed consent. Adipose stromal cells (ASCs) were isolated from adipose tissue using guidelines specifically approved by Biomedicine Ethics Committee of Institute of Biomedical Problems, Russian Academy of Sciences (Physiology Section of the Russian Bioethics Committee, Russian Federation National Commission for UNESCO, Permit #314/MCK/09/03/13), as previously described [29] with some modifications [30]. Briefly, tissue samples were treated with 0.075% collagenase IA (Sigma-Aldrich, USA). After washing, cells were resuspended in αMEM (Gibco, USA), supplemented with 10% FBS (Hyclone, USA), 250 μg/ml amphotericin, 5μg/ml streptomycin, 5U/ml penicillin, and 2mM glutamine (MP Biomedicals, USA).
Mononuclear cells from umbilical cord blood (cbMNCs) were isolated after written informed consent in the Cord blood bank "Cryocenter" (Moscow, Russia) using guidelines of the License of Federal Service on Surveillance in Healthcare and Social Development (Roszdravnadzor) (Permit #FS#2010/342). Cryopreserved samples of cbMNCs were provided as a part of Scientific Agreement between Cryocenter and Institute of Biomedical Problems.
The study was approved by Biomedicine Ethics Committee of Institute of Biomedical Problems of Russian Academy of Sciences.

ASC/cbMNCs, ASC/cbHSPCs and ASC/cbHSPCs-1 co-culture
To enrich the population of cbHSPCs we used the experimental approach described by us previously [28]. Briefly, confluent (70-80%) ASC layers were pre-formed in 35 mm Petri dishes at 20% and 5% O 2 . Then, cbMNCs (1x10 6 /ml) were seeded on these feeders in RPMI 1640 (Gibco, USA), supplemented with 10% inactivated FBS (Hyclone, USA), 250 μg/ml amphotericin, 5 μg/ml streptomycin, 5 U/ml penicillin, and 2mM glutamine (MP Biomedicals, USA) and co-cultured at the same O 2 concentrations as ASCs. After 72 hours (Day 3), all floating cbMNCs were carefully washed away. One Petri dish in each set was fixed with cold methanol and the attached cbMNCs were analysed after Giemsa staining of the cocultures. The other dishes with attached cbHSPCs were further expanded with replenishment of the RPMI 1640 culture medium twice per week. After 96 hours (Day 7) of ASC/ cbHSPCs co-culturing, the generation of new non-adherent cell suspension (cbHSPCs-1) was observed. Non-adherent cbHSPCs-1 were harvested, enumerated, and assayed by flow cytometry, seeded in colony-forming cell assays and reseeded on new ASC feeder.
To evaluate the potential of newly-formed cbHSPCs-1, 40x10 3 cells/ml were transferred onto new pre-formed ASC layers, again at 20% and 5% O 2 accordingly. After 72 hours (Day 11), the morphological observation of attached and floating cbHSPCs-1 was performed, nonadherent cells were removed and co-cultures were further expanded until Day 14.

Colony-forming cell (CFC) assay
Floating cbHSPCs-1 were collected after 72 h of co-culture with ASCs and 50x10 3 cells/ml were plated in methylcellulose-based medium MethoCult H4534 (Stemcell Technologies, USA) according to the manufacturer's protocol at 20% and 5% O 2 . Colonies were scored after 14 days.

Cobblestone Area-Forming Cells (CAFCs)
CAFCs were defined as clusters of small, tightly packed cells that were non-refractory when viewed under a phase contrast microscope.

MultiPlex Flow Cytomix Assay
Chemokine concentration in conditioned medium (CM) after ASC/cbHSPCs co-culture was measured using Human Chemokine 6plex (G-CSF, IL-8, MCP-1, MIG, MIP-1α, MIP-1β) assay (Bender MedSystems, Austria) on FAX Calibur flow cytometer (BD, USA). CELLquest software (BD) was used for data acquisition. The concentration of each cytokine was linearly dependent on fluorescence intensity and was calculated using Standard curves that were generated for each cytokine using Flow Cytomix Pro software (eBioscience, USA).

Light and scanning electron microscopy (SEM)
Bright-field, phase contrast and NAMC analysis of cultured cells was performed using Nikon Eclipse Ti-U microscope equipped with a Colour Digital Camera DS-Ri1. Images were saved and later processed with NIS-Elements Auto Research software (Nikon Instruments, Japan). For SEM ASCs and ASC/cbHSPCs were prepared routinely: cells were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer for 1 h at room temperature, post-fixed in OsO 4 (1% aqueous solution on 0.1 M phosphate buffer, 1 h at room temperature), and dehydrated in ascending acetone concentrations. Dehydrated specimens were dried at a critical point dryer in liquid CO 2 in Polaron device (Great Britain) and sputtered with gold (100 nm thick) using ion sputtering technique in Eiko device (Japan). Examination of samples was performed using a Hitachi S-500 scanning electron microscope (Japan), equipped with capture self-made device at accelerating voltage of 25 kV with following resolution: 512X512 pixels, 8 bit/pixel.

Statistical analysis
All data were derived from at least three independent experiments. The results are presented as mean ± standard error of the mean (M±S.E.M). Comparisons between experimental results were determined by Mann-Whitney test for independent samples; p<0.05 was considered statistically significant.

Characterisation of cbHSPCs, attached to ASCs
On Day 3 mature blood-borne cells and myeloid progenitors of different commitments were detected among attached cells with differential Giemsa staining: promonocyte/monocytes, erythrocytes; lymphocytes, neutrophilic and eosinophilic myelocytes, neutrophilic and eosinophilic metamyelocytes (Fig 1A-1G). Besides, cbHSPCs with a basophilic cytoplasm and a large nucleus with 1-2 nucleoli without specific lineage features were demonstrated, which made it possible to refer to them as non-differentiated haematopoietic progenitors (blasts) (Fig 1G).
The adhered cbHSPCs preserved high viability, as shown by flow cytometry after Annexin V-FITC/PI staining. (Table 1).
After a further 72-96 hours (Day 7), the single and clustered cbHSPCs were revealed on the ASC surface (Fig 2A and 2B). Besides the adhered cbHSPCs, floating cells were also detected in the co-culture (Fig 2C and 2D). As described previously, these cells are the progeny of the adhered HSPCs proliferating on the stromal layer [31]. These floating cbHSPCs (cbHSPCs-1) were collected and their functional activity was further evaluated. Some cbHSPCs migrated into the space under ASCs, forming cobble-stone like structures, whose area enlarged throughout further culturing (Fig 2E and 2F).

Characterisation of the newly-formed cbHSPCs-1
Number and viability. The number of cbHSPCs-1 was similar at 20% and 5% O 2 and made up 140±84 and 115±50x10 3 /ml, respectively. The trypan blue exclusion test demonstrated about 90% viable cells both at 20% and 5% O 2 .
Immunophenotype. Over 95% of cbHSPCs-1 cells were bearing CD45 pan-leukocyte antigens. The share of CD34 + cells among HSPCs-1 significantly increased in comparison with unmanipulated cbMNCs and was higher at 5% O 2 (p<0.05). Moreover, we failed to identify CD133 + cells in cbMNCs before expansion, but about 20% of HSPCs-1 were CD133 positive both at 20% and 5% O 2 . The enrichment of HSPCs-1 suspension with undifferentiated CD34 + progenitors as a result of expansion on ASCs made up about 6 and 8 times at 20% and 5% O 2 , respectively (Fig 3A and 3B).
Evaluation of CFCs among cbHSPCs-1. To reveal CFCs, HSPCs-1 were grown in methylcellulose-based media for colony-forming unit (CFU) assays (MethoCult H4434) for 14 days and total CFC and lineage-restricted CFC numbers were calculated. The total CFU number of HSPCs-1 at 20% O 2 , was almost 1.6 times lower in comparison with at 5% O 2 (Fig 3C(i)). More than half of the CFUs were represented by BFU-E, whose numbers were twice as high at 5% O 2 than at 20% O 2 . Besides, there were more CFU-G and fewer CFU-M at 5% O 2 (Fig 3C(ii)).  cbHSPC-1 reseeding. cbHSPCs-1 suspension was plated on the mitomycin C-treated ASC layer at 20% and at 5% O 2 . On Day 11, a floating population of cells was again formed above the attached cbHSPCs-1 (Fig 4A and 4B), single adhered cbHSPCs-1 and newlyformed clusters could be seen, similar to those described above, on the surface of ASCs ( Fig  4C and 4D). Besides, after 2 weeks of co-culture, we observed large CAFC-occupied areas under ASCs at both 20% and 5% O 2 , similar to that described above for unmanipulated cbHSPCs (Fig 4E and 4F).

Paracrine regulation of ASC/cbHSPCs interaction
To characterise the contribution of paracrine regulation into ASC/cbHSPC interactions, we evaluated the production of 6 chemokines (G-CSF, IL-8, MCP-1, MIG, MIP-1α, MIP-1β) in the conditioned medium (CM) of ASCs monoculture and ASC/HSPCs-1 co-culture with Human Chemokine 6 Plex. Only MCP-1 and IL-8 were detected in CM from the ASC monoculture, with the concentration of both being higher at 20% O 2 (Fig 5A). In co-culture, CM was collected when newly generated floating cbHSPCs-1 were harvested on Day 7 (96 h after washing away the unmanipulated cbMNCs). Co-culture was accompanied with considerable increase of the concentrations of both chemokines at 20% as well as at 5% O 2 (Fig 5B), in addition, the production of MIP-1β was detected (Fig 5C).

Discussion
Cord blood MNCs are a heterogeneous population of blood-borne cells with the vast majority of mature blood-borne elements and a small subpopulation of cbHSPC progenitors of different commitment. In this paper, we used the "physiological" approach to enrich cbHSPC fractions, as described earlier [28]. The total cbMNCs were co-cultured with ASC feeder for 3 days, after which all of the non-attached cells were removed. It was shown previously that the stroma-adhered HSPCs were the most rapidly proliferating and least committed [12,32]. Among the attached cbHSPCs, we detected both mature blood-borne cells and HSPCs of different commitments. Upon further co-culturing of attached cells, mature MNCs disappeared bit by bit, and the population of cbHSPCs was increased due to the appearance of blasts. Besides, the population of floating CD45 + haematopoietic cells appeared and increased above the stromal sub-layer, with their numbers being similar at 5% and 20% O 2 . In addition, we detected CAFClike areas after 2 weeks of co-culturing. These observations are in agreement with those described previously [12,31,32]. In these papers, the three compartments of HSPC distribution in respect to the stromal layer for ex vivo haematopoiesis were demonstrated: adhered to the stroma surface, below the stroma and floating cells.
At present, mainly immunoselected CD34 + cells from the bone marrow, cord blood and mobilised peripheral blood MNCs are implicated for HSPC expansion [3,10]. This approach has some drawbacks. The CD34 + population is very small and loss during immunoselection extraction may be fatal [33]. Additional experimental procedures along immunoselection may result in damage of a part of the selected cells. The selection of only CD34 + cells leaves primitive progenitors with the CD34 -/CD133 + phenotype in the unmanipulated MNC suspension. There is one more very important point. Some papers revealed the uncertainty as to whether all HSCs/HPCs express CD34 [34] and whether CD34 expression by human stem cells is a reversible process [27,35]. Moreover, Koller et al. described an enrichment of the immature populations after culturing BM MNCs, suggesting that complete CD34 + selection might not be strictly necessary to obtain net cell expansion [36]. Lastly, successful expansion of CD34 + cells requires the addition of exogenous cytokines [3,10]. In this connection, our approach for HSPC enrichment at the expense of adhesion of poorly differentiated haematopoietic cells on ASCs and their further expansion in the conventional growth medium without additional cytokine supplement seems to be very promising tool.
Despite the fact that such feature of haematopoietic niche as low O 2 is well recognised and is considered to be one of the most important mechanisms of haematopoiesis regulation, ex vivo hypoxic protocols are used quite rarely [1,23,24,37,38]. Nevertheless, most of them demonstrated a positive effect of hypoxia on ex vivo haematopoiesis despite the considerable differences in the study design.
Thus, evaluation of haematopoietic HPSC expansion (total cell number, CFU and CD34 + cells) in the presence of microencapsulated osteoblasts revealed a predominant increase in the estimated parameters under 5% O 2 [24]. In Hammoud et al. [25] expansion of CD34 + cells from cord blood on human bone marrow MSCs at 20, 5 and 1.5% O 2 was studied. The authors analysed only floating cells, that are, those that comply with our HSPCs-1. "Hypoxic" cells were most efficient while repopulating the bone marrow of immune-deficient mice in vivo. The authors supposed that the combination of the stromal feeder and low O 2 for ex vivo expansion makes it possible to fully preserve and even improve the functional activity of HSPCs. According to the authors, this is confirmed by the data from CFU-analysis, when the "hypoxic" HSPCs were most abundant with CFCs. In our work, we obtained the same results in different experimental design.
Jing et al. [26] studied the effect of 0.5% O 2 on the interaction of CD34 + cells from mobilised peripheral blood and human bone marrow MSCs. Earlier, the formation of three different compartments of CD34 + cells during 7 days of co-culture was described in detail in this lab: (1) non-adhered fraction, (2) cells that adhered to MSCs which were shining in phasecontrast (the authors called them phase-bright cells) and (3) cells under MSCs that are dark in phase contrast (phase-dim cells). The authors showed that CD34 + cells in these compartments are in a different functional state. Among the adhered cells, most of the proliferating ones can be found, while the majority of resting cells (CD34 + /CD38 -) are found under MSCs. After division, CD34 + cells become unfixed from MSCs and form a floating fraction. In our work, we described the same dynamics and distribution of cbHSPCs-1 (progeny of cbMNCs) on MSCs from adipose tissue. Moreover, despite the fact that in the work by Jing et al. [32] CD34 + cells that migrated under MSCs were not associated with the cells that could form "cobble-stone areas", it is quite evident that these are the same CAFCs that was described earlier by Os et al. [39] and also was demonstrated in present paper. Moreover, we managed to show that floating cbHSPCs-1 represent a heterogeneous population which (if re-cultured), may again yield the cells that occupy all three of the described compartments. Back to the results of Jing et al. [26], the authors managed to show that the area under MSCs at 20% O 2 is the poorest in O 2 . The authors speculate that this contributes to maintaining of the non-committed state of the HSPCs found there. At 0.5% O 2 , the adhesion of CD34 + cells to MSCs was lower, and transmigration under MSCs was higher than at 20% O 2 , which is supposed to have occurred at the expense of the increase in MSC VEGFα production using the HIF-mediated mechanism [26].
MSCs are known to produce a whole range of mediators that take part in regulation of haematopoiesis [12,40]. Among the chemokines involved in this process, special attention is paid to C-C (MCP-1, -2, -3, MIP1-α, MIP1-β, RANTES) and C-X-C (PF-4, IL-8, IP-10) chemokine families that exhibit suppressive activity against myeloid precursors. It was shown that IL-8, PF-4, IP-10, MCP-1 suppressed the proliferation of CFU-GEMM, CFU-GM and BFU-E in vitro, stimulated by growth factors [41]. Here, we demonstrated that in ASC monoculture MCP-1 and IL-8 concentration in CM was higher at 20% O 2 . ASC/HSPCs interaction was accompanied with a considerable increase in MCP-1 and IL-8 production, supposedly due to an increase in chemokine-synthetic activity of ASCs. The increase of both chemokines was more pronounced at 5% O 2 . The functional activity of MCP-1 and IL-8 in ASC/cbHSPCs co-culture is related to inhibition of primitive HSPC cycling, which supports the pool of non-differentiated progenitors [42,43]., Another member of the C-C family-MIP-1β, was revealed in CM of co-cultured ASC/HSPCs. This chemokine does not directly stimulate or inhibit HSPCs; however, being in excess compared with the cognate chemokine MIP-1α, this blocks MIP-1α-mediated inhibition of the proliferation of myeloid progenitors [44]. So, interactions between ASCs and cbHSPCs entails simultaneous increase in cytokines providing negative and positive regulation of self-renewal and differentiation of cbHSPCs, which generally ensures a balanced paracrine profile that supports both primitive and committed haematopoietic progenitors.
Thus, the combination of unfractionated cbMNCs and human ASCs enables the selection of functionally active cbHSPCs, the further expansion of which without exogenous cytokines provides enrichment with CD34 + cells. ASCs efficiently support the viability of cord blood haematopoietic progenitors of different commitment at standard and tissue-related O 2 levels at the expense of direct and paracrine cell-to-cell interactions. A reduced O 2 concentration makes it possible to increase the share of CD34 + cells in the cbHSPC population and ensure the predominant development of certain haematopoietic lineages. These data are of importance both from the viewpoint of fundamental physiological haematopoiesis mechanisms and in connection with the importance of directed ex vivo cbHSPC expansion for the needs of regenerative medicine.