The plasma levels of soluble HLA-G molecules correlate directly with CD34+ cell concentration and HLA-G 14bp insertion/insertion polymorphism in cord blood donors

Abstract

Background

Cord blood provides haematopoietic stem cells for allogeneic transplantation and, thanks to the naivety of its immune system, has several advantages over other sources of stem cells. In the transplantation setting, the presence of immunosuppressive human leucocyte antigen (HLA)-G molecules has been advocated to prevent both rejection and Graft-versus-Host disease. HLA-G is physiologically expressed throughout pregnancy and is contained in cord blood at birth. Moreover, it has recently been reported that not only cord blood mesenchymal cells, but also CD34⁺ cell progenies produce soluble HLA-G (sHLA-G). We tried to identify the largest producer of sHLA-G among 85 healthy cord blood donors at Pavia Cord Blood Bank, correlating the sHLA-G concentration with the HLA-G 14bp insertion/deletion (INS/DEL) genotype and CD34⁺ cell concentration.

Materials and methods

We measured sHLA-G levels in 36 cord blood plasma stored at −20 °C for 2 months and 49 cord blood plasma stored at −196 °C for 4–6 years, by enzyme-linked immunosorbent assay. All cord blood donors were genotyped for the HLA-G 14bp INS/DEL polymorphism by polymerase chain reaction. For each cord blood unit, we measured the cell concentration by flow cytometry.

Results

We did not find differences in sHLA-G levels between cord blood plasma aliquots stored for 4–6 years at −196 °C and cord blood plasma aliquots stored for 2 months at −20 °C. We observed a higher sHLA-G concentration in cord blood plasma donors who carried the HLA-G 14bp INS/INS genotype and had higher CD34⁺ cell concentrations (P =0.006).

Discussion

This is the first report showing that the best cord blood stem cell donor is also the best sHLA-G producer, particularly if genetically characterized by the HLA-G 14bp INS/INS genotype. If the therapeutic role of sHLA-G molecules were to be finally established in the transplantation setting, our data suggest that cord blood plasma donors can provide a safe source of allogeneic sHLA-G immunosuppressive molecules ready for transfusion.

Introduction

Over the years, cord blood (CB) has become a valid source of haematopoietic stem cells for allogeneic transplantation worldwide^1,2. Compared to other stem cell sources, CB has the intrinsic properties of lowering the incidence of Graft-versus-Host disease (GvHD) and enabling tolerance of higher degrees of human leucocyte antigen (HLA) mismatches between donors and recipients³. These benefits mainly derive from the fact that CB is a product of pregnancy and maintains the same characteristics of the foetal immune system, such as its naivety.

During pregnancy, several immune mechanisms are triggered to induce tolerance between mother and foetus; in particular the expression of human leucocyte antigen-G (HLA-G) molecules on cytotrophoblasts can block the maternal attack by inhibiting uterine natural killer (NK) cell lysis and CD8⁺ T-cell cytotoxicity, suppressing allogeneic CD4⁺ T-cell proliferation and promoting T helper 2 activation^4–7.

From an immunological point of view, as half the foetal genetic patrimony derives from the father, the foetus must be considered a semi-allograft to the mother, and the pregnancy-related HLA-G molecules serve to down-regulate the maternal immune system allowing implantation-engraftment. In solid organ transplants, it has been shown that high levels of soluble HLA-G (sHLA-G) in the blood are correlated with reduced incidences of acute and chronic graft rejection in heart and kidney recipients^8–12. For this reason, sHLA-G levels have been proposed as a way of non-invasive monitoring for organ transplanted patients during their entire follow-up¹³. Moreover, high sHLA-G blood levels have been demonstrated to prevent both rejection and GvHD in allogeneic peripheral blood stem cell transplantation¹⁴.

Taking into account that CB derives from pregnancy and HLA-G molecules act in both pregnancy and transplantation as immune-modulators, we aimed to identify the best sHLA-G producers among our CB donors. First, we investigated whether the pregnancy-related levels of sHLA-G were preserved in CB units also after medium-term cryostorage. We then tried to correlate the sHLA-G plasma concentrations with the genotype of the corresponding donors, considering the HLA-G 14bp insertion/deletion (INS/DEL) polymorphism, which is characterized by an insertion of 14 nucleotides in the 3′UTR of exon 8 affecting mRNA stability and protein expression¹⁵. Finally, as it has been recently reported that, besides CB mesenchymal stem cells, CD34⁺ cell progeny produce HLA-G, we correlated the sHLA-G levels with both total nucleated cell concentration and CD34⁺ cell concentration to improve the classification of our CB donors as sHLA-G producers.

Materials and methods

Study subjects

Our study included 85 healthy Caucasian neonates contributing to the Pavia Cord Blood Bank (IRCCS Policlinico San Matteo Foundation, Pavia, Italy). All mothers signed written informed consent to voluntary participation in the CB banking programme for unrelated stem cell transplantation. According to FACT Netcord international standards and national laws, a complete medical history of the neonate’s family, including mother, father, siblings and grandparents, was obtained to exclude genetic and infectious diseases^16,17. An exhaustive obstetric history (i.e. of previous pregnancies) was also required as women with recurrent spontaneous abortions are ineligible to donate CB. Finally, during each pregnancy, the health of both the mother and the foetus was carefully monitored, in order to prevent the storage of stem cells derived from pathological pregnancies, including preterm ones (<37 gestational weeks).

Cord blood collection and total nucleated cell count

CB was collected immediately after vaginal or Caesarean delivery from umbilical cord vessels by the in utero technique, using a CB collection kit with 20 mL of citrate-phosphate-dextrose (CPD) anticoagulant solution (Fresenius KABI, Bad Homburg, Germany). Before cryopreservation, the volume of the CB was reduced by plasma depletion¹⁸. The CB plasma was discarded except two aliquots (each of 2 ml), stored at −196 °C together with the CB unit for future testing. The total nucleated cell (TNC) count was determined by an automated haematology analyser (CELL-DYN Sapphire, Abbott Laboratories, Abbott Park, Illinois, United States of America). Only CB units containing ≥1,200×10⁶ TNC after manipulation were banked. Starting from the TNC count, we corrected the TNC concentration by the dilution factor of the anticoagulant (20 mL of CPD).

Analysis of haematopoietic CD34⁺ cell populations by flow cytometry

The population of haematopoietic progenitor cells was measured by flow cytometry. Cells were stained with CD34 (phycoerythrin) and CD45 (fluorescein isothiocyanate) antibodies (BD Biosciences, San José, CA, USA), and the CD34+ cell count was determined using both single- and double-platform methods, according to the guidelines of the International Society of Haematotherapy and Graft Engineering¹⁹. Starting from the CD34⁺ cell count, we corrected the stem cell concentration by the dilution factor of the anticoagulant (20 mL of CPD).

Molecular analysis

Genomic DNA was extracted from the CB and maternal peripheral blood (collected at the time of delivery) using an automated DNA extraction system supported by the AGOWA^® mag Maxi DNA Isolation Kit PLUS (AGOWA GmbH, Berlin, Germany).

We genotyped 85 healthy infant donors for the HLA-G 14bp polymorphism by polymerase chain reaction (PCR) analysis using specific primers, as described by Tripathi and colleagues²⁰. According to the HLA-G 14bp genotyping, the CB units were subdivided into three groups: HLA-G 14bp INS/INS, INS/DEL and DEL/DEL.

Measurement of soluble HLA-G G1 and G5 isoforms by enzyme-linked immunosorbent assay

The levels of soluble HLA-G (sHLA-G) in CB plasma were determined using the sHLA-G enzyme-linked immunosorbent assay kit (Exbio, Praha, Czech Republic) for the detection of soluble G5 and shed transmembrane G1 molecules. The assay was performed following the manufacturer’s instructions. All samples were run in duplicate and mean absorbance was determined at a wavelength of 450 nm for each sample. A calibration curve was constructed by plotting absorbance values against concentrations of calibrators, and sHLA-G concentrations were determined according to this standard curve. Haemolysed or lipaemic samples were excluded from analysis. We corrected the sHLA-G concentration by the dilution factor of the anticoagulant (20 mL of CPD).

Statistical analysis

The median and interquartile range (IQR) were used to summarize quantitative variables as they were not normally distributed (Shapiro’s test). The Mann-Whitney test for independent data (for comparisons between two groups) or Kruskall-Wallis ANOVA (for comparisons among more than two groups) with Bonferroni’s correction was used to analyse differences in sHLA-G levels. Pearson’s r coefficient was used to evaluate the correlation between parameters. P values <0.05 were considered statistically significant. All tests were two-sided. The data were analysed using the STATA statistical package (release 9.0, 2006, Stata Corporation, College Station, Texas, USA).

Results

We first measured the sHLA-G concentration in 36 CB plasma aliquots stored at −20 °C for 2 months (median value of 34.50 ng/mL, IQR: 12.87–51.36). Next, we determined the sHLA-G concentration in CB plasma samples maintained at − 196 °C: 17 samples stored for 4 years (median value of 52.36 ng/mL, IQR: 32.63–59.61), 20 samples stored for 5 years (median value of 48.05 ng/mL, IQR: 35.08–77.36) and 12 samples stored for 6 years (median value of 36.03 ng/mL, IQR: 13.25–54.87). The sHLA-G levels did not significantly differ among CB plasma samples stored from 4 to 6 years in liquid nitrogen (−196 °C), and even the comparison between short-term storage (2 months) at −20 °C and medium-term storage (4–6 years) at −196 °C did not show significant differences in sHLA-G concentrations.

We then considered all the 85 CB plasma samples to verify the correlation between sHLA-G concentration in CB plasma and the HLA-G 14bp insertion/deletion (INS/DEL) polymorphism genotype of the corresponding infant donors. Twenty-three CB donors carried the HLA-G 14bp INS/INS genotype and had a median sHLA-G plasma concentration of 49.63 ng/mL (IQR: 37.90–72.27), 37 CB donors carried the HLA-G 14bp INS/DEL genotype and had a median sHLA-G plasma concentration of 33.59 ng/mL (IQR: 13.19–52.61), and 25 CB donors carried the HLA-G 14bp DEL/DEL genotype with a median sHLA-G plasma concentration of 42.00 ng/mL (IQR: 25.46–54.14). The highest sHLA-G levels were observed in HLA-G 14bp INS/INS carriers (P =0.005, INS/INS vs INS/DEL).

We also analysed possible correlations between the sHLA-G levels of all 85 CB plasma and the TNC and CD34⁺ cell concentration in the corresponding CB units. We did not find a correlation between TNC and sHLA-G concentrations (r =0.07, P =0.5469). We did, however, find a statistically significant trend that directly correlated the levels of sHLA-G with CD34⁺ cell concentration (r =0.2484, P =0.0263): the higher the stem cell concentration the higher the levels of sHLA-G molecules (Figure 1).

Figure 1.

Correlation between sHLA-G levels (ng/mL) and the CD34+ cell concentration (10⁴/mL).

Finally, we verified whether the correlation between the sHLA-G levels in the CB plasma aliquots and the CD34⁺ cell concentration in the corresponding CB unit remained unchanged according to the HLA-G 14bp INS/DEL genotype. We found a statistically significant correlation between sHLA-G and CD34⁺ cell concentrations in the group of HLA-G 14b INS/INS carriers (r =0.5662, P =0.0060, Figure 2), whereas the correlation was not statistically significant in the other two groups of HLA-G 14b INS/DEL and HLA-G 14bp DEL/DEL carriers.

Figure 2.

Correlation between sHLA-G levels and CD34+ cell concentration in the HLA-G 14bp INS/INS carriers.

Discussion

Pregnancy-related HLA-G molecules exert an immunosuppressive action, avoiding foetal rejection by the maternal immune system. The HLA-G class Ib gene expresses seven isoforms derived from alternative splicing of a primary transcript: the membrane-bound molecules G1, G2, G3, G4 and the soluble forms G5, G6, G7. During pregnancy, both transmembrane and soluble HLA-G molecules down-regulate the maternal immune response by binding to several inhibitory receptors, such as KIR2DL4 (on natural killer cells), LILRB1 (on monocytes, dendritic cells, T lymphocytes, B lymphocytes and natural killer cells) and LILRB2 (on myeloid-derived cells)^5–7.

In the setting of kidney and kidney/pancreas transplants, measurement of pre- and post-transplantation sHLA-G levels showed that lower concentrations were present in patients with signs of rejection in biopsy specimens^13,14. In agreement with these data, high sHLA-G blood levels were found in heart recipients with a reduced incidence of acute and chronic graft rejection^8–10. Some authors have, therefore, proposed monitoring soluble sHLA-G levels during the entire follow-up after solid organ transplantation¹³. If the infusion of allogeneic sHLA-G molecules were to be found to improve solid organ engraftment in recipients producing low levels of sHLA-G, then a suitable donor would need to be identified.

At present, there are reports that CB CD34⁺ cells can be cryopreserved and maintained for 15–20 years in liquid nitrogen for transplantation purposes, based on data collected by the New York CB Bank, the oldest bank in the world²¹. To our knowledge, besides the quality controls analysing stem cell viability and function in cryopreserved CB, there are no data on the maintenance over the years of pregnancy-related molecules enhancing tolerance between mother and foetus, such as sHLA-G. We, therefore, investigated whether sHLA-G molecules are preserved after cryostorage considering the oldest CB plasma aliquots stored in our CB bank in Pavia. We compared sHLA-G levels among CB plasma samples stored at − 196 °C for 4 to 6 years (17 CB samples for 4 years, 20 CB samples for 5 years, and 12 CB for 6 years) and also between 36 CB plasma samples stored at − 20 °C (for 2 months) and 49 CB plasma samples stored at −196 °C (for years), and found no differences. Our results suggest that sHLA-G is present and stable in CB plasma also after medium-term cryostorage. However, as the Pavia CB bank started to deplete plasma from 2003, a larger set of data including older CB units should be considered to verify the presence of sHLA-G molecules also after long-term storage (more than 10 years).

Moving onwards, trying to identify sHLA-G “high producers” by a molecular technique (PCR) simpler than immunohistochemistry, we correlated the HLA-G 14bp INS/DEL genotypes of 85 CB donors with the corresponding sHLA-G plasma levels.

There are conflicting data in the literature about the correlation between HLA-G 14bp genotype and HLA-G mRNA levels in pregnancy. In 2003, two different studies were published on this issue: on the one hand, it was observed that the HLA-G 14bp INS mRNA isoform was expressed at a lower level than the 14bp DEL mRNA in 12 first trimester placentas; on the other hand, the HLA-G 14bp INS mRNA isoform, which generates an additional splice site removing 92 bases from exon 8, was reported to be more stable than the HLA-G 14bp DEL mRNA isoform (complete mRNA form) by radiolabeled reverse transcriptase PCR and quantification in placenta cell cultures^15,22.

Among our 85 CB donors, we found the highest sHLA-G plasma levels in HLA-G 14bp INS/INS carriers (P =0.005, INS/INS vs INS/DEL). Our findings are in line with data recently reported by Gonzalez and co-workers, who observed a direct correlation between the HLA-G 14bp INS allele and higher HLA-G5 plasma levels in pregnant women²³. However, as there is no unanimous consensus about which HLA-G 14bp allele produces more HLA-G proteins, further studies need to be done to confirm that HLA-G 14bp INS/INS carriers are the best sHLA-G plasma producers among CB donors.

In CB transplantation, the cell dose, in terms of both TNC and CD34⁺ cell count, is known to be a major prognostic factor affecting the outcome. When choosing a donor it is, therefore, of the utmost importance to define the cellular characteristics of each CB unit^24,25. Interestingly, it has been recently reported that, besides mesenchymal stem cells, CD34⁺ cell progeny also produce sHLA-G in CB^26,27. For this reason, we correlated sHLA-G levels with both TNC and CD34⁺ cell concentrations in order to improve the classification of our 85 CB donors as sHLA-G producers. We observed a direct association between sHLA-G levels and CD34⁺ cell concentration (P =0.0263, r= 0.2484), but not between sHLA-G and TNC (P =0.5469, r =0.0663, Figure 1). Furthermore, we correlated sHLA-G levels with the cellular characteristics according to the HLA-G 14bp INS/DEL genotype of the corresponding CB donor, confirming the direct association between sHLA-G and CD34⁺ cell concentration in the group of HLA-G 14bp INS/INS carriers (P =0.0060, r =0.5662, Figure 2). Thus, according to the findings in the CB units, the best CB donor is also the best sHLA-G donor.

Plasma depletion is routinely used in CB banking to optimise the cryostorage of CB units and reduce costs. Although this space-saving policy must be maintained, the waste of CB plasma implies the loss of the pregnancy-related proteins, such as sHLA-G. If the immunotolerogenic potential of allogeneic sHLA-G in the transplantation setting is proven, separate storage of CB plasma aliquots could provide a precious and safe source of tolerogenic molecules, which may be infused as an adjuvant to the conventional immunosuppressive therapy²⁸.