The ratio of ADSCs to HSC-progenitors in adipose tissue derived SVF may provide the key to predict the outcome of stem-cell therapy

Patients applied to California stem cell treatment centers (Cell Surgical Network) and all participants signed an IRB-approved consent form. Liposuction aspirates were obtained under local anesthesia using an IRB-approved protocol (International Cell Surgical Society; IRB# ICSS-2016-024) from donors as previously explained in details [14]. In short, patients received local anesthesia consisting of lidocaine 0.5% with epinephrine. They underwent liposuction procedure utilizing 2.5–3 mm cannula.

To assess subjective outcomes, standard questionnaires and evaluation system for follow-ups were utilized. Details about which questionnaires were used were published previously [14]. Data were collected via e-mail and telephone and entered into a customized TrackVia (Denver, Colorado) database. All responses were voluntary and patients did not receive compensation to participate. To assess the potential benefit of the treatment, a benefit index by grading the improvements on a scale 1–5 based on the patient reports was generated. Each of these numbers was assigned by the principal investigators based on review of the following subjective outcomes tests for each condition. For category#1 (orthopedic conditions), the following tests were administered to patients: Visual analog pain scale plus, knee injury and osteoarthritis outcome score, WOMAC Score, Hip injury and osteoarthritis outcome score, disability arm shoulder hand score, foot and ankle outcomes questionnaire, oswestry back questionnaire, neck disability index. For category #2 (organ and tissue dysfunction including neurologic disease, cardiac disease and urologic disease), the following tests were administered for each specific condition: AQoL-4, cardiac status form, Minnesota heart failure questionnaire, erectile hardness grading form EHGS, Peyronie’s bother score, pelvic urgency frequency form and O’Leary-Sant form. For Category 3 (Autoimmune function), the following tests were administered: autoimmune status form and AQoL-4.

Stromal vascular fraction cells were isolated following established protocols [14]. 50 cubic centimeter (cc) of adipose tissue was collected into a TP-101 syringe (single use sterile fat processing syringe) and condensed by centrifugation at 2800 rpm for 3 min in the Time Machine centrifuge. Good manufacturing practices (GMP) grade Collagenase (Roche GMP grade, Heidelberg, Germany) containing 12.5 Wünsch units was added and the lipoaspirate was placed on a Time Machine shaker/incubator in 37 °C for 30 min to enzymatically digest the collagen matrix in order to procure the SVF inside closed Time Machine syringes (TP-102 syringe by MediKhan, Los Angeles, USA) in the operating room. Collagenase from Roche is assayed in Wünsch units. One (Collagen Degrading Units) CDU catalyzes the hydrolysis of 1 μmol of l-leucine equivalents from collagen in 5 h at 37 °C, pH 7.4 and 1 Wünsch unit = ~ 1000 CDU. The Time Machine™ (USA trade name for the MediKhan Lipokit system—Los Angeles, CA, USA; 510(k) approved for fat grafting) is a specialized closed surgical processing system for isolating stromal vascular fraction. It combines a centrifuge and incubator to prepare SVF in a closed system. The end product of 5–10 ml SVF was collected using a sterile 20 ml Luer Lock syringe (Sigma-Aldrich, St Louis, MO, USA). The patient’s SVF is then injected back into their body, either directly into the inflamed area (such as the knee or hip joint) or into the bloodstream, via an intravenous injection. SVF deployments in most orthopedic cases were I.V and intra-articular [14].

Immunophenotyping, cell count and cell viability of freshly isolated SVFs were performed by flow cytometry on a 4-color FACSCalibur flow cytometer (Becton–Dickinson). Removal of erythrocytes was carried out by using RBC lysis buffer (GIBCO). The following monoclonal antibodies (MAbs) conjugated to various fluorochromes were used in single cell suspensions of SVFs: CD45 (clone HI30), CD34 (clone 8G12), CD90 (clone 5E10), CD31 (clone WM59), CD73 (clone AD2), CD14 (clone M5E2), CD206 (clone 19.2), CD105 (clone SN6), CD44 (clone L178), CD66b (G10F5), CD235a (GA-R2 (HIR2)), CD3 (clone HIT3a), CD19 (clone HIB19), CD56 (clone B159). All antibodies were purchased from BD Biosciences, except the anti-human CD105 (clone SN6), which was purchased from e-biosciences, and staining was performed according to standard staining protocol as described before [34]. Cell viability was verified by 7-amino-actinomycin D (7-AAD) or Propidium Iodide (PI) and Annexin-V staining (BD Biosciences).

A portion of freshly isolated SVF cells was plated in a CELLstart CTS (Gibco)-coated plate with culture media. The serum-free culture media was composed of DMEM (Invitrogen), 5% Stemulate (Cook Regentec, IN, USA), 1 × GlutaMAX (Invitrogen), MEM non-essential amino acids, and Pen/Strep (Invitrogen). Stemulate is a GMP grade pooled human platelet lysate, which supports the expansion of multiple types of cells including MSCs. It has essential growth factors and other proteins such as platelet-derived growth factors (PDGF-AA, PDGF-AB/BB), epidermal growth factor (EGF), vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), transforming growth factor beta 1 (TGF-B1), fibrinogen, regulated upon activation, normal T -cell expressed, and secreted (RANTES), and chemokine ligand 5 (CCL5), C-X-C Motif Chemokine Ligand CXCL1/2/3 that are necessary for cell growth and functions. Primary cells were cultured for 7–14 days and were defined as “Passage 0” (P0). The medium was replaced every 3 days, and cells were passaged every week or two when 80% confluence was reached. 1 × TrypLE Express (Life Technologies) was used to dissociate cells. Immunophenotyping on different passages were carried out as described above.

The ability of cells to differentiate into mesodermal lineages was tested in adipogenic or osteogenic differentiation mediums provided by iXcells Biotechnologies. Differentiation was initiated by seeding MSC at a density of 20,000 cells per cm². After 1–2 weeks of culture in defined media conditions, cells were stained with Oil Red O (Sigma) for the detection of lipids and Osteoimage Mineralization Dye (Lonza) to visualize osteogenic mineralization. The quantification of intracellular lipid droplets was performed using AdipoRed Assay kit from Lonza, following the instructions provided by the manufacturer. Fluorescence measurement was carried out at Tecan Infinite 200 plate reader with excitation at 485 nm and emission at 572 nm. For the quantification of mineralization, Osteoimage Mineralization Assay was utilized (Lonza). Excitation and emission wavelengths were 492 and 520, respectively.

Cells were processed and incubated with the following antibodies, CD45, CD235a, CD34, CD31 and CD146, as described above. The viability dye PI was added before sorting for exclusion of dead and apoptotic cells. Cell sorting was performed using on a FACSAria III cell sorter (BD Biosciences) equipped with a class I biosafety cabinet. Four-way sorting was performed at 10,000-to-20,000 events per minute. Sorted populations were placed into flat bottom 48-well plates coated either with fibronectin (Millipore) or CELLstart CTS (Gibco). DMEM + 5% Stemulate culture media was used to culture sorted cells. Viability was determined by Trypan blue exclusion on day 5–7 post-culture.

The data were analyzed by comparing a “benefit index” of the treatment to various patient factors such as age, gender, body mass index (BMI), dosage, and route of injection (further described in the proceeding sections). Table 1 contains the raw benefit index numbers for each patient. The ratios of various cell subtypes were also calculated, and compared to the various patient factors in order to determine which, if any, factors were correlated to treatment efficacy.

Most of the human clinical studies regarding implantation of autologous adipose SVF containing ADSCs have specifically assessed effectiveness and safety, without identifying the factors that might be responsible for potential benefits. Therefore, in this study we analyzed patient’s characteristics and SVF prepared by using standard procedures in order to assess the post-treatment clinical outcome. The autologous SVF was returned back to the donor patients with various disease conditions such as orthopedic, inflammatory, degenerative tissue or organ, and autoimmune diseases within 2–3 h. All patients underwent treatment with SVF cells as scheduled and no complications related to adipose tissue processing and SVF cells preparation was noticed. There were no reported side effects associated with SVF cell therapy. We followed the patients every 3–4 months for about a year post-treatment, to assess the potential benefit of the treatment. All patients were requested to fill out a questionnaire, and we generated a benefit index by grading the improvements on a scale 1–5 based on the patient reports.

Grade numbering:

5-Complete recovery

4-A large difference (A noticeable & continuous difference)

3-Some improvements

2-Minimal improvement

1-No improvement.

Aiming to identify clinical factors predictive of response to SVF therapy, we first asked whether gender of the patients played a role in the effectiveness of treatment. Table 1 list all of the patient’s gender, age, medical reason, delivery methods and benefit index. Out of 24 females, 4 patients did not participate in the follow-up program (N/A), and 5 patients reported no benefits (25%). Furthermore, 1 patient with “Complete recovery” (5%); 6 patients with “A big difference” (30%); 5 patients with “Some improvement” (25%), and 3 patients with “Minimal improvements” (15%), were detected respectively. For male patients, we observed almost an identical trend; Among 16 male patients, 12 patients benefited at various degrees (Complete recovery: 0%, A big difference; 31%, Some improvement; 31%” and Minimal improvements; 13%), whereas 4 patients did not report any benefit (25%). A bar graph, in Fig. 1a shows the mean benefit index for gender with 95% confidence interval. The mean benefit index was 2.69 for male versus 2.75 for female. Therefore, we argue that the gender did not play a role in the patients reported therapy outcome. We subsequently analyzed whether the age of the patients was a contributing factor. The treatment age started with the youngest patient who was 16 years old, and the oldest patient who was 91 years old. The majority of patients were between 40 and 79 years old. The largest population of patients, at 15, was in the age group 60–69, followed by 11 patients in the age group of 50–59, 6 for 70–79, and 5 for 40–49. There was only 1 patient for each of the following groups: < 20; 20–29; 30–39 and 90–99. When the ages of all patients were plotted against the benefit index for direct correlation, we could not see any difference for the mean benefit index among the different ages (Fig. 1b). Therefore, we concluded that age could not be a specific determining factor for treatment success. Next, we divided the patients into three main classification categories based on their clinical conditions to investigate whether any particular disease or condition benefitted most from the treatment. All of the orthopedic applications along with joint associated chronic pain, stiffness, deformity, decrease in the range of motion, and arthritis were included in category #1 (Table 1). Organ or tissue dysfunctions were grouped under category #2, and autoimmune diseases were grouped under category #3. We further indicated the route of SVF injection for each patient in our analyses: IV injection alone (I.V.), direct localized injection alone (D.I.), or both. We had the highest number of patients in category #1 with 29, followed by 13 and 2 patients for category #2 and #3, respectively. We analyzed the number of patients within category #1 using the benefit index scale from 1 to 5, and compared the data with the distribution of category #2. Furthermore, we analyzed and compared the route of injections with their different benefit index. The data analyses indicated that neither the specific clinical condition nor the route of injection played a significant role in clinical outcome since patients from both of the categories, #1 and #2, and with different route of injections, could be found in every benefit index to almost the same extent. Taken together, the administration of the autologous SVF for the treatment of different medical conditions has a positive effect, but the factors such as gender, age, clinical condition, and route of injection did not play a decisive role in the treatment efficacy nor are the proof of the true underlying mechanism of improvement.

Next we investigated whether the stem/progenitor cells composition and dose of SVF could show a correlation with treatment efficacy. Initially, we hypothesized that the presence of one of the major cell subsets injected at optimal numbers would favor the observed improvements. Earlier studies reported the heterogeneity in SVF and great variability in the percentages of different subsets [35, 36]. Some reports have focused mostly on the difference in the isolation methods, handling, variability of cell yields and mode of administration [37, 38]. To date, there are insufficient data to establish a correlation between SVF-dose, and direct therapeutic effect. Therefore, we conducted a comprehensive analysis by characterizing the cells types with regenerative potential present in the donor SVF samples. To determine the composition of SVF, we performed flow cytometry analysis and identified 4 distinct cell populations based on the surface expression of CD45 and CD34 (Fig. 2a). Two of these populations which are CD34⁻ CD45^high and CD34⁻CD45^low, were also present in the blood (Fig. 2a1). On the other hand, the CD34⁺ positive subsets: CD34^high CD45⁻ and CD34^low CD45⁺, were only detected in the SVF samples. In 21 out of 44 patients the combined ratio of all 4 populations was almost 100 percent of the nucleated cells. In these patients, the average percentage of CD34⁻ CD45^high, CD34⁻CD45^low, CD34^high CD45⁻, and CD34^low CD45⁺ was 16, 44, 28, 12%, respectively (data not shown). In 18 patients, on the other hand, the percentage of double negatives (DNs; CD34⁻ CD45⁻) showed distinct variations in three different ranges; 5–15% (n = 8), 30–70% (n = 2), and 80–99% (n = 13). Figure 2a panel 2 through 5 shows a representative analysis of CD34 and CD45 for each of these groups. The variation in the presence of fibroblasts, pre-adipocytes, smooth muscle, mature endothelial cells and the remaining RBCs following the lysis is believed to be the main contributor for the differences of DNs among patients.

We further investigated SVF samples, in more detail using cell type specific surface markers in patient samples. There is not a unique marker for CD34⁺ cell subsets, therefore a combination of markers were used for more comprehensive analyses. Figure 2b shows the phenotype of the subsets based on flow cytometry analysis. CD31 (PECAM-1) is a classic marker for endothelial cells and their progenitors. In combination with CD146, we distinguished CD34^high CD45⁻ CD31⁺CD146⁺ AT-ECs separately from CD34^high CD45⁻CD31⁻ CD146⁻ ADSCs within the CD34^high cell subset. The fraction of EC compartment among CD34^high subset displayed a great variability by ranging from 1 to 55% with an average of 21%, across all of the donors with DN < 2. On the other hand, CD34^low CD45⁺ cells, co-expressed CD206 and CD14. This subset has been described as Hematopoietic/Monocyte-like progenitor cells (HSC-progenitors) in the literature [39‐41]. Back-gating analysis of this subset on forward and side scatter (FSC and SSC) dot plot revealed that they reside on top of ADSCs indicating they are more granular (Fig. 1c). Heterogeneous expression of CD105 and CD73 was detected among SVF samples. Figure 1b shows a representative of well-separated subsets within the ADSCs; CD34⁺CD105⁺ CD73⁻ and CD34⁺105⁻CD73⁺ cells. The percentage of pericytes described as CD45⁻CD34⁻CD31⁻CD146⁺ cells made up only a small fraction (average percentage of nearly 1%) of whole SVF. CD45⁺ cells were equally abundant in SVF, and the majority of these cells were CD66b⁺ Neutrophils (data not shown). Other observed immune cell populations included B-cells, NK cells, and T-cells and it is highly likely they were derived from blood vessels co-harvested during the procedure. The viability of CD34⁺ cells was verified regularly by 7-AAD/PI and Annexin-V staining and they were 80–90% viable post-isolation.

In conclusion, the following four main adipose-resident subsets were identified in SVF samples: ADSCs: CD34^high CD45⁻ CD31⁻ CD146⁻, HSC-progenitors: CD34^low CD45⁺ CD206⁺CD31⁻ CD146⁻, AT-ECs: CD34^high CD45⁻ CD31⁺CD146⁺, and Pericytes: CD45⁻CD34⁻CD31⁻CD146⁺. We believe that the abundance of one or some of these subsets in combination reflecting some level of orchestration were the primary factors which determine the treatment outcome.

Few previous studies compared some of the patient factors such as age, gender, BMI that might affect the yield, composition, purity and potency of the freshly isolated SVF samples, reported controversial results [35, 42‐45]. To determine whether the patient’s disease conditions may alter the SVF’s yield and composition, we also included 14 cancer patients. Enumeration analyses lead us to evaluate the possible effect of tumor growth on the cell composition of SVF and further assess the feasibility of utilizing SVF in cancer patient’s treatment. Identical liposuction procedure was followed for such patients and samples were analyzed for the presence of the same 4 subsets of cells. Due to the variations in the ratio of double negatives among all of the patients (non-cancer and cancer), we calculated the absolute number of isolated cell subsets as shown in Table 2 and then performed a correlation analysis between patient characteristic and the SVF yield. We first checked the gender difference for a possible factor in harvested stem cells; ADSCs and HSC-progenitors. We found that the average absolute numbers of ADSCs in females (5.6E+3/per cc of fat/non-cancer) was almost threefold higher than the average yield in male patients (1.9E+3/per cc of fat/non-cancer) (Fig. 3a). A similar difference was observed with the absolute number of HSC-progenitors: 3E+3/cc of fat of HSC-progenitors in females versus 1.2E+3 cells, in males. Regarding the absolute number of AT-ECs, the difference between males and females was fivefold: 1E+3/cc ECs in females versus 212E+0 cells, in males. No difference was observed for pericytes. Interestingly, a similar trend was observed in between 8 male and 6 female cancer patients, although all of the 4 cell subsets were overall lower (at least two–threefold) in number compared to non-cancer patients (Fig. 3b). The average absolute number of ADSCs, HSC-progenitors, AT-ECs and pericytes in female versus male cancer patients were 1.9E+3 versus 948E+0, 1.9E+3 versus 682E+0, 417.5E+0 versus 70.3E+0, 75.3E+0 versus 56.3E+0, respectively. Furthermore, different age categories in 9-year increments were made. In females, the treatment age range was 16–82 years old and marked the end points of eight different age groups. There were no patients in the age group of 20–29 and a single patient in each of these groups: 10–19, 30–39 and 80–89. Therefore, the yield of 4 SVF subsets were calculated in four different groups, each with more than 3 patients. The average total number of ADSCs was 5.9E+3 in the age group between 40 and 49, reached its highest, 10.3E+3, in the age group 50–59, and declined in the subsequent groups: 3.7E+3 (60–69); 4.5E+3 (70–79). For HSC-progenitors, the highest average total number was recorded in the same group 8.5E+3 (50–59) but remained almost the same (1.5E+3 and 1.6E+3) in the other age groups. In AT-ECs and pericytes, the distribution of average count per age showed the following patterns: 658E+0/(40–49), 1.8E+3/(50–59), 700E+0/(60–69), 1.7E+3/(70–79), and 82E+0/(40–49), 165E+0/(50–59), 158E+0/(60–69), 93E+0/(70–79), respectively. On the other hand, in males treatment ages started very late, around the age of 50–59 and only two more age groups (60–69 and 70–79) could be used for statistical analysis due to the presence of only one patient in other groups (80–89 and 90–99). A continuous decline of the actual number was detected in all of the cell subsets except pericytes among these age groups; male ADSCs 2.4E+3 (50–59); 1.8E+3 (60–69), 1.2E+3 (70–79); male HSC-progenitors 1.6E+3 (50–59); 1.1E+3 (60–69), 700E+0 (70–79); male AT-ECs 368E+0 (50–59); 143E+0 (60–69), 128E+0 (70–79); male pericytes 171E+0 (50–59); 172E+0 (60–69), 51E+0 (70–79). All of the cancer patients were also included in the analysis and no statistically significant difference was observed among 4 age groups as shown in the Fig. 3d. The average BMI for all of the participants was 25.5. The BMI was compared for two groups: BMI ≤ 25 and BMI > 25. Lean persons were defined with BMI < 25 [46]. The average BMI for the group BMI ≤ 25 was 21.3 and for BMI > 25 it was 29.7. Although it was not significant, the average total number for ADSCs and HSC-progenitors were higher in BMI > 25 group when compared to BMI ≤ 25 group. The average numbers for BMI > 25 groups were ADSCs: 4.6E+3, HSC-progenitors: 3.1E+3, AT-ECs: 716.6E+0, and pericytes 145.2E+0. On the other hand, the average numbers for 25 ≤ BMI groups were ADSCs: 3.2E+3, HSC-progenitors: 1.3E+3, AT-ECs: 612.5E+0, and pericytes 127.5E+0 (Data not shown). Over all, Fig. 3 summarizes the correlation of absolute number, with the gender, and age of the donors in the cancer versus non-cancer patients. In conclusion, the donor characteristics of gender, to some extent age, and BMI of the adipose tissue donor, and their physiological condition such as cancer or non-cancer, have the potential to impact the total number of isolated adipose-specific cells. These findings are similar to those obtained by most of the other groups [44, 45].

The majority of the SVF samples were used for treatment in the clinics and a relatively minor fraction was sent to us for further analysis. When there were enough samples, a portion of the cells sent for characterization was also used for cell culture analysis. Cellular expansion of adherent cells was initiated in a culture medium supplied with 5% Stemulate and adipose-derived and expanded-mesenchymal stem cells (ADE-MSCs) cultures were further analyzed for morphology, phenotype, and differentiation. These cells exhibited the expected fibroblast-like morphology following their adherence to the plastic (Fig. 4a) and continuous culture resulted in phenotypically homogeneous population as periodically verified by flow analysis (Fig. 4b). Phenotypic characterization of cultured cells revealed that all of the samples shared the same markers, which have been described, by International Federation of Adipose Therapeutics and Sciences (IFATS) and the International Society for Cell Therapy (ISCT) for the minimum criteria for identifying MSCs. They highly expressed CD90, CD73, CD105, and CD44, while remained negative for CD45, CD146, CD31, and CD34 [22]. There was no variation between the phenotypes of cells from different passages or freezing–thawing cycles. The last criterion to identify ADE-MSCs is to test their multipotency by differentiating into different cell types such as osteoblasts, and adipocytes. Therefore, in order to confirm the mesenchymal stemness of expanded cells, their differentiation potential was tested. ADE-MSCs from different donors at 0–1 passages were induced to differentiate into adipocytes or osteoblasts for 8–14 days in defined culture conditions. The cells started to differentiate into adipocytes as early as day 3 and were fully differentiated at day 14 as consistent with increased cell vacuolation of fatty acid observed under the light microscope using Oil O Red staining (Fig. 5a). The following figure (Fig. 5b) shows the partition of fat droplets of differentiated adipocytes as detected by AdipoRed reagent at day 14. The differentiation was further confirmed by quantification of the accumulation of intracellular triglycerides using the same reagent and measuring the fluorescence in four independent experiments. Relative fluorescence unit (RFU) was 14-fold higher in induced cells (5240 ± 1690) compared to non-induced cells (365 ± 233). Culturing the cells under osteogenic condition induced the cells to differentiate within a week. Osteogenic differentiation (Fig. 5a, b) was confirmed by the formation of aggregates and osteoimage mineralization assay. At day 8, the average calcium deposit in osteoblast reached into tenfold difference compared to undifferentiated control (18,241 ± 2480 versus 1800 ± 430) as quantified by RFU on a plate reader using 492 nm excitation and 520 nm emission wavelengths.

Understanding the key features of the cellular components of SVF is vital for appreciation of the potential uses and contributions in tissue maintenance and repair. Therefore, we wanted to know which cell subset(s) gave rise to the expanded cell populations. Stem cell (ADSCs and HSC-progenitors) components as well as EC and pericytes of fresh SVF were simultaneously sorted and individually put into culture to observe expansion. Viable (PI-) single cells (height-to-width ratio) were isolated to purity close to 100% (Fig. 6) and put into 48-well plates coated with either Fibronectin or CELLstart CTS. Among 4 different subsets, only CD34^high CD45⁻ CD31⁻ CD146⁻ cells formed fibroblast-like morphology and started to divide. This led to the conclusion that ADSCs represent an origin of the ADE-MSC grown in culture, a finding that has been validated by other groups [47‐49]. Although neither pericytes nor HSC-progenitors expanded, Trypan blue exclusion analysis determined that 40–60% of the adhered cells were still alive up to 5–7 days post-culture. Next, we asked whether the therapy with expanded cells was as efficient as SVF treatment and evaluated the therapeutic potential of autologous MSC cultured under GMP conditions by American CryoStem. Table 3 summarizes patient’s gender, age, medical condition, the route of injection, injected total number of cells, and patient’s satisfaction assessment. The overall average response rate was 3.1 and did not show a significant difference from the response rate (2.7) observed in SVF treatment group. Therefore we concluded that the therapy with expanded MSC cells originating from CD34⁺ subset was not more efficacious compared to SVF treatment.

Although the patient’s characteristic we followed seemed to affect the SVF yield, they did not show a correlation with the treatment efficacy. We also checked the SVF cell composition and found out that a particular cell subset did not play a dominant role in the positive clinical outcome. In our cell culture experiments, only the ADSC component of SVF could give rise to expanded cell populations. Furthermore, the therapy with expanded MSCs did not suggest more efficacies compared to SVF treatment. The question remains; do other cell subsets directly contribute to tissue repair/regeneration as a result of a complex interactive network? Since all of the patients were given a combination of cell subsets, we asked whether a certain combination rather than individual subset might play a role in determining the treatment efficacy. We first explored the relative ratios of total injected numbers of ADSC: HSC-progenitors: AT-ECs: Pericytes. Since the number of pericytes made the lowest portion among these 4 subsets, we started our analysis by calculating the ratio based on pericytes. However, we couldn’t derive any meaningful conclusion (data not shown). Therefore, the pericytes were excluded and the relative ratios were next calculated based on 3 subsets. Table 4 shows the ratio of ADSCs: HSC-progenitors: AT-ECs, based on total number of each injected cell subset.

The degree of the ratio of ADSCs to HSC-progenitors in the given treatment was associated with overall treatment efficacy, whereas their lower ratio resulted in a higher benefit index on average. The data from Table 4 are summarized in Fig. 7. A downward trend can be seen between the benefit index and the mean ratio as the benefit index increases (Fig. 7a). The error bars represent the 95% confidence interval. The total average of the ratio of ADSCs to HSC-progenitors in patients who benefited highly from the treatment was 1.84, which signifies that every HSC-progenitor should be accompanied by at least two ADSCs to achieve the maximum patient benefit. For the next groups of patients with the benefit indexes of 3 and 2, the average for the ratio became 2:85 and 7.73, respectively. The patients who reported no benefit, and the patients who did not report back to the clinic, both shared a close average ratio value of approximately 4 (3.63 and 4.99, respectively). In another separate representation the raw data from which (A) was derived was plotted for every single patient (Fig. 7b). Downward trend is again clearly visible suggesting a correlation between the ratio and the benefit of the treatment.