Clinical implementation of a procedure to prepare bone marrow cells from cadaveric vertebral bodies

Abstract

Aims

Vertebral bone marrow is a rich and easily accessible source of hematopoietic and mesenchymal stem cells that has been used to promote chimerism and transplantation tolerance in connection with cadaveric organ transplantation. The purpose of this study is to provide a detailed account of the procedure used to prepare the first five vertebral bone marrow products for infusion in conjunction with the first hand/hand–forelimb transplants performed at the University of Pittsburgh (PA, USA).

Materials & methods

The cell separation and release testing were performed at the University of Pittsburgh Cancer Institute’s Hematopoietic Stem Cell Laboratory, a Good Manufacturing Practice-compliant facility accredited for clinical cell processing by the Foundation for Accreditation of Cellular Therapy (FACT) and for clinical flow cytometry by the College of American Pathologists (CAP).

Cadaveric bone marrow is a rich source of viable hematopoietic and mesenchymal stem cells. Bone marrow [1–4] and granulocyte colony-stimulating factor-mobilized peripheral blood [5] have been used as a cell-based approach to promote transplantation tolerance in cadaveric and directed donor solid organ transplantation, respectively. The combination of peritransplant donor bone marrow infusion under the cover of bolus immune suppression has become known as the Pittsburgh Protocol [1–4]. Recently, we have expanded the use of this technique to hand and forelimb grafts [6], where the indication for transplantation is quality of life improvement and there is a pressing need to protect the patient from the risks of lifelong immunosuppression. In order to implement the Pittsburgh Protocol in this setting we preformed preclinical experiments to develop and validate a Good Manufacturing Practice (GMP)-compliant method for preparation of a cellular product from cadaveric vertebral bodies [7]. In this article, we describe the implementation of this methodology as a clinical protocol, and publish the full working procedure as Supplementary Material (see online at www.futuremedicine.com/doi/suppl/10.2217/RME.11.89).

Materials & methods

Vertebral body dissection

Vertebral body (VB) dissection was initiated from beating heart cadaveric organ donors after cross-clamping, organ and hand retrieval. The research protocol was approved by the University of Pittsburgh Committee on Research Involving the Dead. The clinical transplantation protocol was approved by the University of Pittsburgh Internal Review Board. A total of nine vertebrae (Th8–L4) were harvested according to our published procedure [7]. VBs were rinsed with a saline–betadine solution and double-bagged with 2 l of sterile Custodiol^® preservation solution (Essential Pharmaceuticals, LLC, Newtown, PA, USA). The bagged VB were labeled according to protocol and stored in a cooler on wet ice for transportation. VB were held on ice for a mean time of 37 ± 18 h (mean ± standard deviation [SD]) prior to initiation of processing.

Isolation of bone marrow cells

Isolation of bone marrow cells (BMC) was performed at the University of Pittsburgh Medical Center Hematopoietic Stem Cell Laboratory, a Foundation for the Accreditation of Cellular Therapy (FACT)-accredited GMP compliant facility with equipped with a class 10,000 clean room. The translational studies performed to develop and validate this procedure have been documented [7]. The full working procedure formatted according to the recommendations of the Clinical and Laboratory Standards Institute (CLSI) [8] is available as Supplementary Material and includes reagents and supplies, instrumentation, reagent preparation instructions, cautionary notes, detailed methods, acceptable end points, result reporting and quality control sections. Briefly, VBs were surface decontaminated by immersion in hydrogen peroxide (3% in normal saline) and cut into 1–2 cm cubes. Bone cubes were irrigated with medium consisting of albumin (0.5 g/dl), DNAse (350 U/ml), MgCl₂ (2.5 mM), sodium heparin (10 U/ml) and gentamicin (50 μg/ml), and crushed in a prototype instrument designed to crush bone cubes (bone marrow grinder; Biorep^® Technology, Miami, FL, USA) housed in a class II biosafety cabinet (Figure 1A). Crushed bone was serially strained through stainless steel sieves (425 and 180 μm), rinsed with medium, and the cell-rich filtrate was collected. Rinsed bone fragments were resuspended in 250 ml of medium and tumbled for 20 min using a prototype bone marrow tumbler (Biorep Technology) (Figure 1B). After tumbling the bone fragments were sieved and the cell-rich filtrate was pooled. The tumbling steps were repeated. All cells collected from a total of three washes were pooled and filtered (500 and 200 μm) using a bone marrow collection filter set. The filtered product was transferred in 600-ml transfer pack bags and centrifuged at 700 × g at ambient temperature, and resuspended to 140 ml in saline, 5% human serum albumin.

Figure 1. Apparatus for preparation of bone marrow cells from vertebral bodies.

(A) Cubed vertebrae are fed into the bone crusher, which renders it into a bony paste. The crushed bone is continuously irrigated with processing medium containing heparin and DNAse. The bone crusher is operated in a class II biosafety cabinet located inside the class 10,000 clean room. (B) The bone tumbler rotates and gently agitates the crushed bone, releasing bone marrow cells into the processing medium.

Cell counts

Nucleated cell counts were obtained using a COULTER^® AcT diff2^™ Hematology Analyzer (Beckman Coulter, Inc., Brea, CA, USA), according to the manufacturer’s instructions. These results were used to calculate recovery of nucleated cells during processing and are approximate, since analyzers using particle impedance to detect objects tend to overestimate the number of nucleated cells in bone marrow due to the presence of cell-sized fat globules and other acellular debris. Final cell dosing calculations were based on absolute CD45 counts determined by flow cytometry.

Flow cytometric enumeration of CD45⁺ & CD34⁺ cells & viability determination

Flow cytometry was performed using a commercially available single platform lyse/no wash assay (Stem-Kit^™, Beckman Coulter, Inc.) according the manufacturer’s instructions. Stem-Kit measures absolute numbers of CD45⁺ and CD34⁺ cells by comparison of the number of measured events to that of an internal bead standard of known concentration. Viability of CD45⁺ and CD45⁺/CD34⁺ cells was measured by exclusion of 7-aminoactinomycin D. Data were acquired on a four-color EPICS^™ XL Cytometer (Beckman Coulter, Inc.), which was calibrated daily with Flow-Check^™ and Flow-Set^™ beads (Beckman Coulter, Inc.). CD34 reference cells (low, normal and high controls; Streck Laboratories, Omaha, NE, USA) were run with each sample as a positive control of known CD34 content. Data were acquired immediately and analyzed in real time using System II software (Beckman Coulter, Inc.) and a modification of the Sutherland method [9,10]. The University of Pittsburgh Medical Center Hematopoietic Stem Cell Laboratory, in which the flow cytometry studies were performed, is accredited by the College of American Pathologists (CAP) for clinical flow cytometry.

Sterility cultures & cryopreservation

Aerobic, anaerobic and fungal cultures were performed as previously described [7]. Cells were suspended in 10% dimethyl sulfoxide (DMSO) to a concentration of 4 × 10⁸ CD45⁺ cells/ml and cryopreservation was performed using a controlled-rate freezer (CryoMed^®, Thermo Scientific, Waltham, MA, USA) as previously described [10]. Cells were held in liquid phase in liquid nitrogen until use.

Infusion

The product was serially thawed and infused at the bedside. Each cryopreserved bag was placed inside a sterile thawing bag and immersed in sterile water at 37°C in a heated water bath. The thawed product was inspected for the presence of visible cell clumps and was aseptically aspirated into a 60 cc syringe. If cell clumps were observed, the syringe was connected to a 500 μ bone marrow filter. The cell suspension was pushed into an intravenous access port or central line over a period of 3–5 min. After the entire product (usually four cryobags) was infused, the line was flushed with saline. Patients were monitored for untoward effects for 4 h after completion of infusion.

Viability of the thawed product

A viability sample (approximately 1 ml) was held from the last syringe, dispensed into a cryovial, held on ice and transported back to the laboratory (mean time to assay: 48 ± 4 min, mean ± SD). The product was diluted 1:100 in phosphate-buffered saline, to which 100 μl of acridine orange/ethidium bromide working solution (6 and 12 μg/ml in phosphate-buffered saline, respectively) was added. The cell suspension was loaded on a hemacytometer and viability was determined by fluorescence microscopy (Laborlux S, epi-illuminated, I3 cube; excitation: 450–490-nm band pass; emission: 510-nm short-pass emission, 515-nm long-pass suppression; Leitz, Wetzlar, Germany). Viable cells stain green with ethidium bromide, whereas dead cells stain orange with acridine orange [11]. Results were expressed as percentage viable.

Results

Demographics & cell enumeration

Processing time, donor demographics, cell yields and viabilities are shown in Tables 1 & 2. Neither cell recovery nor viability was associated with total processing time, which ranged from 12 to 56 h including specimen hold time (time from VB harvest to initiation of processing). The actual processing itself (time from VB accession in the laboratory to the last cell count) averaged 5 h, exclusive of cryopreservation. The wash and concentration step immediately prior to cryopreservation was accompanied by a variable (32–50%) but significant cell loss, which was occasionally accompanied by visible cell clumping (Table 1). On average, 3.9 ± 0.7 × 10¹⁰ (mean ± SD) CD45⁺ leukocytes were recovered per product after washing and concentration.

Table 1.

Processing time, donor demographics and cell yields.

Case	Hold time (h:min)	Process time (h:min)	Total time (h:min)	Donor age	Donor height (cm)	Donor sex	Donor IBW (kg)	WBC (prewash × 1010)	WBC (precryo × 1010)	Wash recovery (%)	Final flow WBC (× 1010)
HT1	43:47	8:15	52:02	18	188	M	82	11.0	5.0	49.8	4.4
HT2	47:04	5:38	52:42	23	190	M	84	8.5	5.7	67.7	4.6
HT3	6:25	5:50	12:15	50	178	M	73	7.3	4.7	64.1	4.0
HT4	37:46	5:11	42:57	38	160	F	52	5.2	3.5	68.1	2.8
HT5	51:46	4:32	56:18	42	173	F	64	8.1	5.3	65.4	3.8
Mean	37:21	5:53	43:14	34	178		71	7.8	4.9	63.0	3.9
SD	18:01	1:24	18:00	13	12		13	1.8	0.8	7.6	0.7

F: Female; Final flow WBC: Flow cytometric WBC based on the absolute CD45⁺ cell count; HT: Hand transplant; IBW: Ideal body weight; M: Male; Precryo: Hematology analyzer count after washing and concentrating cells for cryopreservation; Prewash: Hematology analyzer cell count of pooled cells immediately after final straining though the in-line bone marrow filters; SD: Standard deviation; WBC: Total white blood cells in product.

Table 2.

Recipient demographics, cell doses and viability before and after cryopreservation.

Case	Recipient age (years)	Recipient IBW (kg)	Total CD34 × 108 (viability)	Total CD3 × 109 (viability)	Precryo CD34 dose 106 cells/kg	Precryo CD3 dose 106 cells/kg	Post-thaw AO viability (%)
HT1	25	78	1.7 (96.2%)	1.6 (36.5%)	2.2	20.0	15.6
HT2	57	73	7.5 (95.4%)	1.4 (46.3%)	10.3	19.7	16.2
HT3	42	82	3.5 (99.5%)	3.3 (98.1%)	4.3	39.6	23.3
HT4	28	57	3.4 (98.6%)	5.1 (79.9%)	6.0	90.2	16.7
HT5	33	52	2.2 (98.3%)	0.9 (95.9%)	4.9	25.5	95.8†
Mean	37	68	3.7 (97.6%)	2.5 (71.3%)	5.4	37.0	18.0
SD	13	13	2.3 (1.7%)	1.7 (28.4%)	3.1	30.7	3.6

^†Filtered at bedside (500 μ bone marrow filter) owing to clumping during thawing procedure, not included in summary statistics.

AO: Acridine orange; HT: Hand transplant; IBW: Ideal body weight; Precryo: Hematology analyzer count after washing and concentrating cells for cryopreservation; SD: Standard deviation.

CD34 content provides an indication of bone marrow quality in bone marrow aspirates and decreases with hemodilution. In these VB-derived samples, it was uniformly high with an average yield of almost 400 million CD34⁺ cells (98% viable) per product. This translated into an average dose of 5.4 × 10⁶ CD34⁺ cells/kg, well in excess of the number needed to mediate full bone marrow reconstitution in a myeloablated recipient. CD3 content indicates the dose of graft-derived HLA-mismatched T cells in the product. CD3⁺ T cells represented 7.1 ± 6.5% (mean ± SD) of CD45⁺ cells, a proportion comparable to what we have previously reported for bone marrow aspirates (9.3 ± 3.5%) used in match sibling bone marrow transplantation [12]. T-cell viability (71%) was considerably lower than CD34⁺ cell viability.

Release criteria

Release criteria for the products were: culture negative (aerobic, anaerobic and fungal), precryopreservation CD34 dose ≥2 × 10⁶/kg, precryopreservation CD34 viability ≥85%. All products met cell dose and viability criteria. A single product appeared to be lightly contaminated with the skin organism Propionibacterium acnes. A confirmatory culture tested negative. The product was approved for infusion under antibiotic coverage according to our institutional protocol.

Cell infusion

Cells were thawed at bedside and infused intravenously. Total cell viability was measured in thawed products by fluorescence microscopy and acridine orange exclusion. As expected, total cell viability was low (16–23%) in thawed samples (Table 2), reflecting the fact that cells of the granulocytic series, particularly those subjected to stress, do not cryopreserve well. Flow cytometric determination of CD34 viability was not performed in the thawed products because the presence of DMSO interferes with the dye uptake assay. In two recipients, the presence of visible cell clumps upon thawing necessitated filtration (500 μ) at bedside. Recipients were monitored for 4 h for acute symptoms associated with intravenous infusion of DMSO-containing cellular products. Only mild reactions were encountered and no unexpected adverse events occurred in association with bone marrow infusion.

Discussion

The impetus to develop a GMP procedure for the isolation of BMC from vertebral bodies was to reduce or eliminate the need for lifelong immunosuppression in hand/hand–forelimb transplant recipients. An independent review of the published literature concluded that infusion of donor bone marrow results in microchimerism in the majority of patients, with effects on rejection episodes that are time dependent and vary with the transplanted organ [13]. The clinical results of the hand transplant recipients whose bone marrow graft products are described here are greatly encouraging [6] and will provide a stimulus for further investigation.

The present study made use of a bead-calibrated single platform flow cytometric assay for absolute CD34 and CD3 counts, a method unavailable when the pioneering work on cadaveric bone marrow was performed. This validated clinical assay is routinely run with three levels of process controls. As it relies on an internal absolute count standard, it is not subject to the several sources of interference commonly encountered in dual platform assays, where Coulter-principle counting is used for total nucleated cell count and immunofluorescence, E-rosetting or flow cytometry is used to estimate the proportion of positive cells. Our results were consistent with the total nucleated cell count (3 × 10⁹/VB) and T-cell content (5–6%) reported by Sharp et al. [14], but were lower than those of Rybka et al. [15], who reported higher total cell recovery (5.7 × 10¹⁰ per nine VBs) and extraordinarily high CD34 content (2.5 vs 0.9% in the present series). We failed to reproduce the assertion that VB BMC have a lower T-cell content than aspirated bone marrow. The percent of T cells in our series was highly variable and ranged from 2.5 to 18.1% (mean 7.1%) and was only marginally lower than that reported in aspirated bone marrow [12].

Prior to implementation of this protocol we performed six practice cell isolation procedures in which the majority of the initial problems were solved [7]. Sterility of the harvested VB, which was a problem in the first practice runs, was addressed by changing surgical practices during VB recovery, including an antibiotic (gentamicin) in the transport medium and surface-decontaminating VBs with hydrogen peroxide immediately prior to processing. Of the five initial products reported here, four were culture negative for aerobic, anaerobic and fungal cultures. One product initially tested positive for a skin organism but was negative upon confirmatory culture, indicating either light contamination of the product or contamination during sterility testing. According to our protocol, the patient’s physician was notified and the sterility results were recorded on the physician infusion request form. Upon product infusion, blood cultures were drawn and patient temperatures were recorded and tracked without incident. All products in the present series met pre-established cell recovery and viability release criteria.

Despite the inclusion of DNAse and heparin in the wash medium, cell clumping with attendant cell loss posed a sporadic problem, necessitating filtration of the product at bedside in two cases. Review of the literature revealed that contrary to common practice, pancreatic DNAse requires the presence of both calcium and magnesium for optimal activity [16]. If only magnesium is added, as was the case in this series, DNAse will be active only if sufficient calcium is present as a low level contaminant. Going to the forward, we have added 0.5 mM CaCl₂ DNAse-containing wash medium (Supplementary Material). This approach may also improve the measured viability after cryopreservation. The release of DNA by dead granulocytes rapidly form aggregates of sticky DNA, which trap viable as well as dead cells.

Intravenous infusion of DMSO-containing cryopreserved cells is commonly associated with a variety of transient side effects including nausea, vomiting, flushing, abdominal cramping, dyspnea, headache, chills, chest tightness, mild bradycardia, hypertension fever and hemoglobinuria/hemoglobinemia [17]. Symptoms may be attributable to vasovagal reaction to infusion of the cold product, the presence of degranulating myeloid cells, or to the DMSO itself. Such reactions are expected and are not treated as adverse events unless they last more than 4 h. In the present series, there were no infusion-associated adverse events.

Taken together, we have documented a GMP-compliant procedure for the preparation of BMC from cadaveric vertebral bodies. The methodology required close collaboration between the surgical organ recovery team and the laboratory personnel. In the present application, the product was cryopreserved and infused in order to promote tolerance of cadaveric hand/hand–forelimb grafts. This procedure may also be of use for other clinical applications, including the preparation of low-passage bone marrow-derived mesenchymal stem cell banks.

Executive summary.

• We have adapted a protocol for the preparation of bone marrow cells from cadaveric vertebral bodies to a good manufacturing practices laboratory environment.

• Cadaveric bone marrow cells may be useful for promoting transplantation tolerance and are a rich source of therapeutic mesenchymal stem cells.