Development and validation of a procedure to isolate viable bone marrow cells from the vertebrae of cadaveric organ donors for composite organ grafting

Abstract

Background aims: Donor-derived vertebral bone marrow (BM) has been proposed to promote chimerism in solid organ transplantation with cadaveric organs. Reports of successful weaning from immunosuppression in patients receiving directed donor transplants in combination with donor BM or blood cells and novel peri-transplant immunosuppression has renewed interest in implementing similar protocols with cadaveric organs.

Methods: We performed six pre-clinical full-scale separations to adapt vertebral BM preparations to a good manufacturing practice (GMP) environment. Vertebral bodies L4-T8 were transported to a class 10 000 clean room, cleaned of soft tissue, divided and crushed in a prototype bone grinder. Bone fragments were irrigated with medium containing saline, albumin, DNAse and gentamicin, and strained through stainless steel sieves. Additional cells were eluted after two rounds of agitation using a prototype BM tumbler.

Results: The majority of recovered cells (70.9 ± 14.1%, mean ± SD) were eluted directly from the crushed bone, whereas 22.3% and 5.9% were eluted after the first and second rounds of tumbling, respectively. Cells were pooled and filtered (500, 200 μm) using a BM collection kit. Larger lumbar vertebrae yielded about 1.6 times the cells of thoracic vertebrae. The average product yielded 5.2 ± 1.2 × 10(10) total cells, 6.2 ± 2.2 × 10(8) of which were CD45(+) CD34(+). Viability was 96.6 ± 1.9% and 99.1 ± 0.8%, respectively. Multicolor flow cytometry revealed distinct populations of CD34(+) CD90(+) CD117(dim) hematopoietic stem cells (15.5 ± 7.5% of the CD34 (+) cells) and CD45(-) CD73(+) CD105(+) mesenchymal stromal cells (0.04 ± 0.04% of the total cells).

Conclusions: This procedure can be used to prepare clinical-grade cells suitable for use in human allotransplantation in a GMP environment.

Conflict of interest statement

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Figures

Figure 1

VB processing. The VB were processed in a class 10 000 clean room. Following separation and removal of soft tissue and intervertebral disks, the VB were divided into cubes using a mallet, osteotome and stainless steel ratchet pruner (A). Bone cubes were irrigated with medium as they were fed into the bone grinder (B) and emerged as fine bone fragments that were placed on a sieve set (C) and rinsed with medium. The rinsed fragments were then placed in the BM tumbler with more medium to dislodge remaining cells (D).

Figure 2

Use of plastic pipe to stabilize the cadaver after removal of nine VB. Heavy-duty polyvinylchloride pipe was notched (inset) and inserted to replace the missing VB. The vertebral arches and spinal processes remained in the donor. The use of rubber O-rings (shown) was discontinued.

Figure 3

Total cell recovery from each wash step. The data comprise the cells recovered at each step of the procedure expressed as a proportion of the total cells in the pooled product. Tukey box-plots show median and quartile values. Whiskers show the range exclusive of outliers, circles show individual data points.

Figure 4

CD34 recovery from each vertebral segment. The data are expressed as a proportion of total CD34+ cells recovered in each product. Tukey box-plots show median and quartile values. Whiskers show the range exclusive of outliers (asterisk), circles show individual data points.

Figure 5

Detailed flow cytometric characterization of VB mononuclear cells. Analysis of a representative sample is shown. The analytical gate for each histogram is shown in square brackets. Numerical results are arithmetic means ± SD for all six samples. Removal of sources of artifact (top row): singlets (mean 5 316 530 ± 2 584 685 events acquired), determined by forward scatter pulse analysis, represented 91.9% of all events; 85.3% of all singlets had DNA content ≥ 2N (GT2N) and 99.7% of these had light scatter properties consistent with intact cells (clean scatter). Autofluorescent events were defined using a compound gate of events falling on the diagonal streaks of fluorescence (FL1) versus FL2, FL1 versus FL3 and FL2 versus FL3. Autofluorescent events had high forward angle light scatter (FS) and intermediate side scatter (SS), or intermediate FS and high SS. They represented a mean of 3.3% of cells within the clean scatter gate, and were eliminated from subsequent analyzes. Taken together, 24% of the events represented potential sources of artifact and were removed from the numerator and denominator of subsequent analyzes. Heme CD34 classifiers (middle row) were detected among clean scatter, non-autofluorescent events according to the International Society of Hematotherapy and Graft Engineering (ISHAGE) protocol, where CD34+ hematopoietic progenitor cells (heme CD34) are defined as CD45+ CD34+, with restricted CD45 expression and restricted light scatter properties. Heme CD34 outcomes are further parameters measured on this population. The majority of CD45+ CD34+ cells were CD133+ (67.1%) and CD117+ (80.3%), with a distinct population of CD90+ CD117 dim primitive hematopoietic stem cells (13.7%). A small subset (0.8%) of CD45+ CD34+ cells, 79.8% of which were CD90+, co-expressed mesenchymal markers CD73 and CD105. MSC, classified (bottom row) as CD45− (non-heme), CD73+ CD105+, represented 0.68% of CD45− cells. The majority of MSC (63%) co-expressed CD90. The right-most panel in the bottom row, gated on all non-autofluorescent clean scatter events, shows the cut-point defining CD90+ events.