Donor Sex and Passage Conditions Influence MSC Osteogenic Response in Mineralized Collagen Scaffolds

Abstract

Contemporary tissue engineering efforts often seek to use mesenchymal stem cells (MSCs) due to their multi‐potent potential and ability to generate a pro‐regenerative secretome. While many have reported the influence of matrix environment on MSC osteogenic response, few have investigated the effects of donor and sex. Here, a well‐defined mineralized collagen scaffold is used to study the influence of passage number and donor‐reported sex on MSC proliferation and osteogenic potential. A library of bone marrow and adipose tissue‐derived stem cells from eight donors to examine donor viability in osteogenic capacity in mineralized collagen scaffolds is obtained. MSCs displayed reduced proliferative capacity as a function of passage duration. Further, MSCs showed significant sex‐associated variability in osteogenic capacity. Notably, MSCs from male donors displayed significantly higher cell proliferation while MSCs from female donors displayed significantly higher osteogenic response via increased alkaline phosphate activity, osteoprotegerin release, and mineral formation in vitro. The study highlights the essentiality of including donor‐reported sex as an experimental variable and reporting culture expansion in future studies of biomaterial regenerative potential.

Bone healing can vary as a function of sex, age, or pre‐existing health conditions. Yet over 95% of tissue engineering and regenerative medicine studies using in vitro models fail to report donor sex. Here, a well‐characterized mineralized collagen scaffold is used to investigate shifts in hMSC osteogenic potential as a function of donor and sex.

1. Introduction

There were 178 million bone fractures that occurred globally in 2019.^{[ 1 ]} The clinical gold standard for bone graft substitutes is allografts and autografts,^{[ 2} , ³ , ⁴ , ^{5 ]} with ≈2 million autologous or allogenic/artificial bone grafts used between 1992 and 2007 in the United States.^{[ 6 ]} Both suffer from key limitations. Notably, limited donor tissue and the need to create a secondary injury site for autografts,^{[ 2} , ³ , ⁴ , ^{5 ]} and concerns about inconsistent purification methods, infection or rejection,^{[ 2} , ³ , ⁴ , ⁵ , ⁷ , ^{8 ]} and variability in healing based on both donor and recipient biology for allografts.^{[ 2} , ³ , ⁴ , ^{5 ]} Tissue engineering offers a framework to develop biomaterials for bone regeneration that may address limitations in allograft and autograft approaches. Many contemporary tissue engineering efforts seek to combine exogenous mesenchymal stem cells (MSCs), morphogens, and biomaterials for bone tissue repair.^{[ 9} , ¹⁰ , ¹¹ , ^{12 ]} MSCs are capable of both differentiating to tissue‐specific lineages as well as generating a complex secretome to support the regenerative action of other cells in the wound site. Further, they can be derived from multiple tissues like adipose (fat tissue) and bone marrow.^{[ 13 ]}

Tissue engineering approaches often rely on cells of unknown origin or that lack diversity related to sex, age, and ethnicity.^{[ 14} , ^{15 ]} Indeed, bone healing can be influenced by many factors including pre‐existing health conditions, age, and sex. Further, sex hormones are important during the development of the skeletal system; hormonal differences can impact peak bone mass and microarchitecture of bone tissue^{[ 16} , ¹⁷ , ¹⁸ , ^{19 ]} and may also underlie sex‐associated variation in bone healing following injury.^{[ 20} , ²¹ , ²² , ^{23 ]} Broadly, male patients have more robust bone healing than female patients.^{[ 20} , ^{21 ]} Sex‐associated variability in bone healing is also observed in animal models, with male animals exhibiting more rapid and robust bone healing than female animals.^{[ 24} , ^{25 ]} Yet, a recent comprehensive review surveyed over 300 articles in the biomaterials literature from 2019 that used in vitro cell experiments and found more than 95% of studies using cell lines, and ≈90% of studies using primary cells, failed to provide information regarding donor sex.^{[ 26} , ^{27 ]} Further, potential variability in healing capacity between donors (e.g., age, underlying health conditions) poses additional challenges for developing biomaterial implants to heal bone injuries.^{[ 28} , ²⁹ , ³⁰ , ³¹ , ^{32 ]} Hence, it is essential that regenerative medicine studies consider the sex, ancestry, age, and underlying health conditions of the cells used.

Regenerative medicine technologies often require large numbers of MSCs, motivating studies to understand the confounding effect of in vitro expansion conditions. In vitro monolayer culture typically used for human MSC (hMSC) expansion can significantly influence the phenotype, proliferation, and differentiation capacity of hMSCs.^{[ 33} , ³⁴ , ³⁵ , ³⁶ , ^{37 ]} Further, most studies on the influence of these factors utilize 2D tissue culture plastic for their culture and subsequent analyses which is not representative of the 3D environment these cells would normally experience. Nonetheless, several groups have found that hMSC activity including differentiation capacity, proliferation, and expression profiles is influenced by donor age^{[ 36} , ³⁷ , ³⁸ , ^{39 ]} and passage number.^{[ 37} , ⁴⁰ , ^{41 ]} Taken together, there is sufficient evidence to show variability between culture methods and donor influence hMSC behavior in 2D culture. However, there is a lack of research in 3D biomaterials that include relevant extracellular matrix cues, and differences in hMSC response may be specific to the biomaterial being developed.

Our lab has developed a 3D mineralized collagen scaffold that promotes osteogenesis in vitro and in vivo without the need for exogenous factors. We have shown these scaffolds promote osteogenesis in vitro using cells from animals^{[ 42} , ⁴³ , ⁴⁴ , ^{45 ]} and humans,^{[ 46} , ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ , ⁵⁴ , ⁵⁵ , ^{56 ]} and improve bone healing in vivo using rabbit^{[ 57 ]} and pig^{[ 58} , ^{59 ]} models. Further, these scaffolds have been used as a model system for mechanistic investigations. MSCs cultured on the mineralized collagen scaffolds promote the secretion of osteoprotegerin which reduces osteoclast resorption in vitro^{[ 54} , ⁵⁵ , ^{56 ]} Moreover, our group has used these materials to understand the antimicrobial effects of manuka honey,^{[ 60 ]} the immunomodulatory effects of placenta‐derived tissues,^{[ 45} , ^{49 ]} the crosstalk between immune cells and hMSCs,^{[ 61 ]} and the potential influence of matrix stiffness (≈1000 kPa dry,^{[ 43} , ^{44 ]} ≈30 kPa, hydrated^{[ 50 ]}), pore size (100–200 um^{[ 44} , ⁴⁵ , ^{47 ]}), and pore alignment (isotropic, anisotropic^{[ 47 ]}) on osteogenic response. The comprehensive characterization of these scaffolds and their demonstrated ability to support osteogenesis motivated their use as a model system to study the influence of donor variability on osteogenic response in vitro.

In this manuscript, we compare metabolic activity and osteogenic gene expression patterns for a panel of donor‐derived hMSCs. We first examine the effect of cell passaging conditions (final passage number; total culture length between thawing from cryopreservation to scaffold seeding) on hMSCs from four donors (two male, two female); we report changes in hMSC proliferation and secretion of pro‐osteogenic factors during a short‐term culture period (21 days). We subsequently examine the donor‐ and sex‐associated variation in proliferation, osteogenic gene expression, and soluble factor secretion using a larger group of hMSCs (four male, four female) over a long‐term in vitro experiment (56 days).

2. Result

2.1. Passage Scheme Influence on Metabolic Activity and Cell Number is Donor‐Dependent

Three passage schemes prior to seeding onto mineralized collagen scaffolds were evaluated (p3‐5: thawed at passage 3 and seeded at passage 5, p3‐6: thawed at passage 3 and seeded at passage 6, and p4‐5: thawed at passage 4 and seeded at passage 5) (Figure 1A) using four donors (B: male, C: female, G: male, and F: female). Cell number and metabolic activity were quantified over the course of 21 days (Figure 2A,B). Cell number remained relatively constant for Donors B and C and increased for Donors F and G (Figure 2A). Donor B (male) and C (female) displayed the highest cell number at all time points. Passage number significantly impacted cell number for Donor B (male), C (female), and F (female) at varying time points, while passage number did not significantly impact cell number for Donor G (male). Passage scheme significantly impacted cell number at Day 7 for Donor F (female), at Day 14 for Donor B (male), C (female), and F (female), and at Day 21 for Donor C (female). Passage scheme p3‐6 appeared to be the worst for cell number in all donors; there were too few cells for seeding Donor G (male), resulting in no cell number data, and cell number was lowest in p3‐6 for Donor B (male), C (female), and F (female) at Day 14 and Donor C (female) and F (female) at Day 21. Normalized metabolic activity remained relatively constant for Donor B (male) and C (female) and decreased after Day 7 for Donor F (female) and G (male) (Figure 2B). The passage scheme significantly impacted normalized metabolic activity on Day 7 for Donor G (male), Day 14 for all donors, and Donor B (male), C (female), and F (female) on Day 21. There does not appear to be a trend in normalized metabolic activity with respect to the passage scheme. Due to low cell number, passage p3‐6 has the highest normalized metabolic activity for Donor B (male), C (female), and F (female) on Day 14 and Donor C (female) and F (female) on Day 21.

Figure 1.

Experimental design and overview. A) Experimental overview for passage scheme study. Human mesenchymal stem cells (hMSCs) from four donors (two male, two female) were cultured on mineralized collagen scaffolds after varying passage schemes to compare the influence of passage number and culture length on cell activity. hMSCs were thawed at passage 3 and seeded at passage 6 (p3‐6), thawed at passage 3 and seeded at passage 5 (p3‐5), and thawed at passage 4 and seeded at passage 5 (p4‐5). Cells were subsequently cultured for 21 days and on mineralized collagen scaffolds and their cell proliferation, metabolic activity, and protein secretion were quantified. B) Experimental overview for donor variability study. hMSCs from 8 donors (4 male, 4 female) were cultured using p3‐5 or p4‐5 passage schemes and subsequently cultured for 56 days on mineralized collagen scaffolds to study the influence of donor and sex on osteogenic response. Cell proliferation, metabolic activity, protein secretion, gene expression, and mineral content were quantified.

Figure 2.

The passage scheme significantly impacts cell number, metabolic activity, alkaline phosphatase activity, and osteoprotegerin secretion. Four donors (B, C, F, and G) with three passage methods (p3‐5, p3‐6, and p4‐5) were cultured on mineralized collagen scaffolds for 21 days to evaluate the impact of culture length and passage number on cell activity. A) Cell number was quantified via DNA extraction. The passage scheme significantly impacts cell number in Donor B, C, and F at varying timepoints, but the passage scheme does not significantly impact cell number in Donor G. B) Metabolic activity was quantified via alamarBlue and normalized to cell number. The effect of the passage scheme on normalized metabolic activity is donor‐dependent. Note: Cell number data for p‐36 Donor G at Day 7 and 14 are missing due to too few cells at the start of the experiment. C) Media was pooled at days 9, 15, and 21 and alkaline phosphatase (ALP) activity was quantified using an ALP activity assay and normalized to cell number. Normalized ALP activity is significantly different for all donors except donor G beginning on Day 9. Note: Normalized ALP activity data for p‐36 Donor G on Days 7 and 14 are missing due to a lack of cell number data. D) Media was pooled at days 9, 15, and 21 and osteoprotegerin (OPG) secretion was quantified using an ELISA. OPG secretion is significantly different between passages at all timepoints for all donors except Donor B. ^**Groups with statistically significant differences based on one‐way ANOVA, do not share the same letters (p < 0.05).

2.2. Passage Scheme Significantly Impacts hMSC Osteogenic Response

Next, alkaline phosphatase activity (ALP) and osteoprotegerin (OPG) secretion were quantified over the course of 21 days (Figure 2C,D). While ALP is well‐established as a biomarker of osteogenic potential, we chose OPG, a soluble decoy receptor to RANKL, due to its role in inhibiting osteoclastogenesis and bone resorption; mineralized collagen scaffolds have previously been shown to promote endogenous production of OPG by hMSCs. Normalized ALP activity remained relatively constant for all donors (Figure 2C). The passage scheme significantly impacted ALP activity for all donors at all timepoints, except Donor G (male) on Day 7. The passage scheme significantly impacts OPG secretion for all donors at all timepoints, except Donor B (male) (Figure 2D). Overall, while the passage scheme significantly impacts all metrics, there were no obvious trends in the passage scheme, suggesting that the optimal passage scheme may be donor‐dependent.

2.3. Donor and Sex Significantly Impact the Cell Number and Metabolic Activity of hMSCs

We subsequently evaluated donor variability as a function of donor‐reported sex on cell activity and osteogenic output using eight donors (C, D, F, H‐female and A, B, E, G‐male). We chose to move forward with passage schemes p4‐5 (thawed at passage 4 and seeded at passage 5) and p3‐5 (thawed at passage 3 and seeded at passage 5) (Figure 1B) to keep the final passage number the same. Cell number and metabolic activity were quantified over the course of 56 days in passage schemes p3‐5 (Figure 3A,B) and p4‐5 (Figure S1, Supporting Information). All donors displayed increasing cell numbers with time, with a less dramatic increase in passage scheme p3‐5 (Figure 3A; Figure S1A, Supporting Information). Male cells displayed higher cell numbers after Day 3 in passage scheme p3‐5 (Figure 3A) and after Day 14 in passage scheme p4‐5 (Figure S1A, Supporting Information). Cell number varied significantly between donors after Day 3 in passage scheme p3‐5 (Figure 3A) and after Day 14 in passage scheme p4‐5 (Figure S1A, Supporting Information). We calculated order of magnitude differences between p3‐5 and p4‐5 average cell numbers (Figure S1B, Supporting Information), and we observed that p3‐5 had higher average cell numbers until Day 7 (sans Donor F‐female) and lower average cell numbers following Day 7 for most donors and timepoints. All donors displayed relatively steady normalized metabolic activity in passage schemes p3‐5 and p4‐5 (Figure 3B; Figure S1C, Supporting Information). Female cells displayed significantly higher normalized metabolic activity at two timepoints in p3‐5 (Day 7 and 28) and p4‐5 (Day 3 and 7), and male cells displayed significantly higher normalized metabolic activity at one timepoint in p3‐5 (Day 14) and two timepoints in p4‐5 (Day 7 and 56) (Figure 3B; Figure S1C, Supporting Information). We calculated order of magnitude differences between p3‐5 and p4‐5 average normalized metabolic activity (Figure S1D, Supporting Information) and observed that p3‐5 had higher average normalized metabolic activity than p4‐5 for most donors and timepoints.

Figure 3.

Donor and sex of donor cells significantly impact cell number and normalized metabolic activity. Eight donors (C, D, F, H‐female and A, B, E, G‐male) with passage scheme p3‐5 were cultured on mineralized collagen scaffolds for 56 days. Metabolic activity and cell number were quantified. A) Cell number was quantified via DNA extraction. Male cells have significantly higher cell numbers after Day 3. B) Metabolic activity was quantified via alamarBlue and normalized to cell number. Female cells have significantly higher normalized metabolically active at Day 7 and 28. Male cells have significantly higher normalized metabolically active at Day 14. ^**Groups with statistically significant differences based on two‐way ANOVA, do not share the same letters (p < 0.05). Sex symbol (male: ♂, female: ♀) indicates significance as a function of donor‐reported sex (p < 0.05) at the indicated timepoint.

2.4. Donor and Sex Significantly Impact Alkaline Phosphatase Activity and Osteoprotegerin Secretion in hMSCs

Next, alkaline phosphatase (ALP) activity and osteoprotegerin (OPG) secretion were quantified over the course of 56 days in passage schemes p3‐5 (Figure 4 ) and p4‐5 (Figure S2, Supporting Information). All donors displayed relatively steady normalized ALP activity with time (Figure 4A; Figure S2A, Supporting Information). Female cells had significantly higher normalized ALP activity until Day 9 and male cells had significantly higher normalized ALP activity at Day 56 in p3‐5 (Figure 4A). Female cells had significantly higher normalized ALP activity at all timepoints in p4‐5 (Figure S2A, Supporting Information). We calculated order of magnitude differences between p3‐5 and p4‐5 average normalized ALP activity (Figure S2B, Supporting Information) and observed that p3‐5 cells had lower average normalized ALP activity than p4‐5 for 5 donors (C, D, H‐female, A, B‐male) and higher average normalized ALP activity than p4‐5 for 2 donors (F‐female, G‐male) (Figure S2B, Supporting Information). All donors displayed increasing OPG secretion with time with a less dramatic increase in p3‐5 compared to p4‐5 (Figure 4B; Figure S2C, Supporting Information). There were no differences in OPG secretion at any timepoint in p3‐5 as a function of donor‐reported sex (Figure 4B), and no differences after Day 3 in p4‐5 (Figure S2C, Supporting Information). We also observe no significant donor variability on OPG secretion in p3‐5 (Figure 4B), but we do observe significant donor variability in p4‐5 (Figure S2C, Supporting Information). We calculated order of magnitude differences between p3‐5 and p4‐5 average OPG secretion (Figure S2D, Supporting Information), finding p3‐5 cells had higher average normalized OPG secretion than p4‐5 for 5 donors (D, F, H‐female and A, B‐male) and lower average normalized OPG secretion for 2 donors (C‐female, G‐male) (Figure S2D, Supporting Information).

Figure 4.

Donor and sex significantly impact alkaline phosphatase activity, osteoprotegerin secretion, and mineral content. Eight donors (C, D, F, H‐female and A, B, E, G‐male) with passage scheme p3‐5 were cultured on mineralized collagen scaffolds for 56 days. Alkaline phosphatase (ALP) activity, osteoprotegerin (OPG) secretion, and mineral content were quantified. A) Media was pooled at days 3, 9, 15, 30, and 56, and ALP activity (U mL cell⁻¹) was quantified using an ALP activity assay and normalized to cell number. Female cells have significantly higher normalized ALP activity at Days 3 and 9. Male cells have significantly higher normalized ALP activity at Day 56. B) Media was pooled at days 3, 9, 15, 30, and 56, and OPG secretion was quantified using an ELISA. There are no significant donor or sex differences in OPG secretion. C) Calcium content at Day 56 was quantified using inductively coupled plasma mass spectrometry (ICP‐MS). There are no significant donor or sex differences in calcium content. D) Phosphorous content was quantified on Day 56 using ICP‐MS. There is no significant sex difference and one significant donor difference in phosphorus content. ^**Groups with statistically significant differences based on two‐way ANOVA do not share the same letters (p < 0.05). Sex symbol (male: ♂, female: ♀) indicates significance as a function of donor‐reported sex (p < 0.05) at the indicated timepoint.

2.5. Female Cells Display Significantly More Mineral Deposition Compared to Male Cells

The mineral content of cell‐seeded scaffolds was determined as the weight percent of calcium and phosphorous measured at Day 56 and compared to an unseeded Day 56 control in passage schemes p3‐5 (Figure 4) and p4‐5 (Figure S3, Supporting Information). While a proxy for dynamic remodeling of mineral content over the course of the experiment, this approach provides a direct measurement of osteogenic activity within the scaffold. Considering the influence of donor‐reported sex, we observe no significant differences in calcium or phosphorus content in p3‐5 cells (Figure 4C,D), but significant differences in calcium and phosphorus content in p4‐5 (Figure S3A,B, Supporting Information). Scaffolds seeded with cells from female donors had significantly higher calcium and phosphorus content at Day 56 in p4‐5 (Figure S3A,B, Supporting Information). Calcium content was higher in cell‐seeded scaffolds compared to unseeded scaffolds in both passage schemes (Figure 4C; Figure S3A, Supporting Information), phosphorus content was lower in seeded scaffolds compared to unseeded scaffolds in p3‐5 (Figure 4D), and phosphorus content was unchanged in cell‐seeded scaffolds compared to unseeded scaffolds in p4‐5 (Figure S3B, Supporting Information). We calculated the order of magnitude differences between p3‐5 and p4‐5 average calcium and phosphorus (Figure S3C, Supporting Information) and observed that p3‐5 cells had lower average calcium and phosphorus content than p4‐5 for all donors (Figure S3C, Supporting Information).

2.6. Male Cells Display Significantly Higher Osteogenic and Immunomodulatory Gene Expression Compared to Female Cells

We subsequently quantified expression patterns for osteogenic, immunomodulatory, and angiogenic genes over the course of 56 days using a custom 38‐gene NanoString panel (35 functional genes and three housekeeping genes). We have plotted gene expression at Day 56 in passage schemes p3‐5 (Figure 5 ) and p4‐5 (Figure S4, Supporting Information), and plotted gene expression for all timepoints in p3‐5 and p4‐5 (Figure S5, Supporting Information). Further, we listed all p‐values and identified sex‐based significance for Day 56 NanoString data (Table S3, Supporting Information).

Figure 5.

Donor and sex significantly impact gene expression at Day 56. Eight donors (C, D, F, H‐female and A, B, E, G‐male) with passage scheme p3‐5 were cultured on mineralized collagen scaffolds for 56 days. A custom Nanostring gene expression panel was used to quantify gene expression. A) 18 osteogenic genes were evaluated. Male cells have significantly higher expression of five osteogenic genes. Female cells have significantly higher expression of four osteogenic genes. B) Fourteen immunomodulatory genes were evaluated. Male cells have significantly higher expression of five immunomodulatory genes. Female cells have significantly higher expression of two immunomodulatory genes. C) three angiogenic genes were evaluated. Female cells have significantly higher expression of one angiogenic gene. **Sex symbol (male: ♂, female: ♀) indicates significance as a function of donor‐reported sex (p < 0.05) at the indicated timepoint. Fold change is plotted as the mean ± standard error.

First, we observe that cells from male donors display significantly higher osteogenic gene expression at Day 56 in p3‐5 and p4‐5 (Figure 5A; Figure S4A, Supporting Information). Specifically, 5 out of 18 genes in p3‐5 and 12 out of 18 osteogenic genes in p4‐5 were expressed significantly higher in male cells. Next, we observe that cells from male donors display significantly higher immunomodulatory gene expression at Day 56 in p3‐5 and p4‐5 (Figure 5B; Figure S4C, Supporting Information). Specifically, 5 out of 14 genes in p3‐5 and 8 out of 14 immunomodulatory genes in p4‐5 were significantly higher in male cells. Finally, we observed that cells from female donors display significantly higher angiogenic gene expression in p3‐5 while cells from male donors display significantly higher angiogenic gene expression in p4‐5 cells (Figure 5C; Figure S4E, Supporting Information). We calculated order of magnitude differences between p3‐5 and p4‐5 average osteogenic gene expression (Figure S4B, Supporting Information), average immunomodulatory gene expression (Figure S4D, Supporting Information), and average angiogenic gene expression (Figure S4F, Supporting Information) and observed that most average gene expression values fall within one order of magnitude. There are some instances when the mean fold change in p3‐5 is the opposite sign as the mean fold change in p4‐5. (i.e., negative fold change in p3‐5 and positive fold change in p4‐5, or vice versa). Specifically, there are 12 instances of osteogenic mean fold change (Table S4, Supporting Information), 14 instances of immunomodulatory mean fold change (Table S5, Supporting Information), and 1 instance of angiogenic mean fold change (Table S6, Supporting Information). Donor G has the most instances of this occurring across all three gene categories.

2.7. Male and Female Donors Release Similar Levels of Osteogenic Factors, While Male Cells Release Significantly More Immunomodulatory Factors

We quantified osteogenic, immunomodulatory, angiogenic, and extracellular matrix (ECM) factors secreted over the course of 56 days in p3‐5 using a custom Luminex panel (Figure 6 ). We also characterize the secreted factor present in the media at all timepoints in p3‐5 (Figures S6 and S7, Supporting Information). Considering osteogenic factors, BMP2 was secreted at the highest levels by Day 56 followed by SPARC, OPN, and MMP13 (Figure 6A). Cells from female donors secreted significantly more SPARC compared to cells from male donors at Day 56; however, no other osteogenic factors displayed significant sex‐associated variation (Figure 6A). For immunomodulatory factors, MMP2 was secreted at the highest levels by Day 56 followed by MMP3, IL6, GAL9, IL1B, and IL10 (Figure 6B). Cells from male donors displayed significantly higher immunomodulatory factor secretion, namely GAL9, IL10, IL1b, and MMP3 (Figure 6B). For angiogenic factors, VEGF was secreted at the highest levels by Day 56 followed by ANG (Figure 6C), and for ECM factors, FN was expressed at the highest levels by Day 56 followed by IGFB3 (Figure 6D). Again, cells from male donors displayed significantly higher IGFB3 (ECM) and ANG (angiogenic) secretion compared to female cells (Figure 6C,D). Donor variability existed at Day 56 for all secreted factors, except MMP13 (statistics not shown on the plot).

Figure 6.

Donor and sex significantly impact osteogenic, immunomodulatory, and angiogenic protein secretion at Day 56. Eight donors (C, D, F, H‐female and A, B, E, G‐male) with passage scheme p3‐5 were cultured on mineralized collagen scaffolds for 56 days. A Luminex assay was used to quantify secreted factors. A) Four osteogenic factors were evaluated. Female cells have significantly higher secretion of 1 osteogenic factor. B) Six immunomodulatory factors were evaluated. Male cells have higher secretion of four immunomodulatory factors. C) Two angiogenic factors were evaluated. Male cells have significantly higher secretion of one angiogenic factor. D) Two extracellular matrix (ECM) factors were evaluated. Male cells have significantly higher secretion of 1 ECM factor. **Sex symbol (male: ♂, female: ♀) indicates significance as a function of donor‐reported sex (p < 0.05) at the indicated timepoint. Cumulative cytokines secreted are plotted as the mean ± standard error.

3. Discussion

We report the influence of the passage scheme on human mesenchymal stem cell (hMSC) response in 3D mineralized collagen scaffolds under development for craniofacial bone regeneration applications. Mesenchymal stem cells are used extensively in research across multiple disciplines for the development of novel therapeutics, tissue engineering solutions, and the understanding of drug efficacy.^{[ 62 ]} Extensive work has been done to study how culture length and freeze‐thaw cycles influence MSC metabolic activity on 2D substrates; however, little research exists about the impact of MSC passage scheme on key indicators of MSC viability and activity in 3D. We tested the influence of time between cell thawing and seeding and the final passage number on hMSC cell response over a 21‐day culture period (Figure 1A), comparing three culture conditions cells thawed at passage 3 and seeded at passage 6 (p3‐6), cells thawed at passage 3 and seeded at passage 5 (p3‐5), and cells thawed at passage 4 and seeded at passage 5 (p4‐5). We observed that the passage scheme had a significant impact on cell number, normalized metabolic activity, normalized ALP activity, and normalized OPG secretion of hMSCs in a 3D biomaterial (Figure 2). This could be due to the final passage number at the time of seeding (passage 6 vs passage 5) or the time cultured on 2D polystyrene before seeding, both of which have been shown to impact hMSC activity in vitro and in vivo.^{[ 49} , ^{63 ]} These data suggest that the passage scheme has a significant impact on cell outcomes and should be considered carefully when developing tissue engineering solutions. Further, these data highlight a critical need to report detailed culture methods including cell passage at thawing and cell passage at experimental use in all future studies.

Beyond the influence of the passage scheme, a more significant parameter that is largely absent and not reported in the literature is the sex of donor cells.^{[ 26} , ⁶⁴ , ^{65 ]} Few in vitro studies have specifically been designed to probe sex‐based differences in osteogenic response. Aksu et al. found that adipose‐derived MSCs displayed greater osteogenic differentiation compared to female cells in vitro on 2D cultures.^{[ 66 ]} However, these studies were done in osteogenic media which contains growth factors and supplements meant to fast‐track the differentiation process. Scibetta et al., showed that muscle‐derived male hMSCs had greater osteogenic potential compared to female hMSCs using in vitro pellet cultures before utilizing an in vivo injury model in mice,^{[ 67 ]} but pellet cultures do not adequately mimic the native 3D bone extracellular matrix. In this manuscript, we evaluated donor variability and potential shifts as a function of donor‐reported sex for hMSCs seeded in our mineralized collagen scaffolds by quantifying their activity and their osteogenic, immunomodulatory, and angiogenic potential. We acquired cells from eight donors (four male, four female) to conduct this donor variability study using both adipose and bone marrow tissue (Table 1 ). Both cell types have shown similar promise in clinical applications.^{[ 68 ]} While adipose cells have been reported to possess higher proliferation while bone marrow cells have been reported to possess higher osteogenic capacity both merit in including them in this study.^{[ 69} , ^{70 ]} We needed to use multiple companies to acquire this large donor set, due both to limited availability per company and the limited availability of female donors for many companies. This illuminates a critical need for expanding the availability of cells from a variety of donors. Nonetheless, we compared these 8 donors cultured in p3‐5 (thawed at passage 3 and seeded at passage 5) and p4‐5 (thawed at passage 4 and seeded at passage 5) passage schemes and seeded them on mineralized collagen scaffolds for 56 days and quantified cell metabolic activity, proliferation, growth factor secretion, ALP activity, and gene expression.

Table 1.

Donor information, company and tissue source, and lot numbers.

Donor ID	Sex	Age	Ethnicity	Pre‐existing conditions	Tissue origin	Company	Lot number
A	M	25	Hispanic	Hypertension Cirrhosis due to ethanol Pancreatitis due to ethanol	Adipose	Essent Biologics	206436
B	M	23	Caucasian	Depression	Adipose	Essent Biologics	206079
C	F	28	Caucasian	Seizure disorder Kidney stones Depression Anxiety PTSD	Adipose	Essent Biologics	205992
D	F	26	Caucasian	Depression ACL, MCL, and meniscus repairs	Adipose	Essent Biologics	211375
E	M	25	Caucasian	None noted	Adipose	Essent Biologics	100069
F	F	20	African American	Not Available	Marrow	RoosterBio	310263
G	M	19	Eritrean/ East African	Not Available	Marrow	RoosterBio	301267
H	F	29	Hispanic	Not Available	Marrow	RoosterBio	00228

We first evaluated the impact of donor and donor‐reported sex on cell metabolic activity and proliferation (Figure 3). In both passage schemes, we observed cells from male donors displaying greater cell proliferation compared to female cells. These differences occurred more often in passage scheme p4‐5 further indicating that passage scheme has an impact on cell metabolic activity and cell number, and this is consistent with our passage scheme experiments (Figure 2). Normalized metabolic activity was observed to be significantly higher in female cells at early timepoints compared to male cells. Previous transcriptomic analysis of human ASCs has observed variations as a function of donor‐reported sex suggesting potential influences on proliferation and differentiation with female cells having a lower potential than male cells^{[ 71 ]} however, others showed limited significance of sex on cell proliferation.^{[ 72 ]} Others have reported that cells from female donors are more capable of counteracting environmental stressors or metabolic disturbances and have a higher resistance to apoptosis compared to male cells^{[ 73 ]} while others have reported metabolic differences on a cellular level.^{[ 74} , ^{75 ]} Taken together, we can conclude that the cohort of cells from male donors we tested was more metabolically active (at later timepoints) and proliferative than cells from female donors on our mineralized collagen scaffolds. This would be an important consideration when using our scaffolds in patients, as the primary challenge in female patients would appear to be cell recruitment and proliferation; however, these might not be a challenge in male patients.

We then examined the donor variability, as well as potential sex‐associated variability, in osteogenic potential via a combination of functional markers (ALP activity; OPG secretion; mineral deposition) and gene expression analysis. ALP is highly expressed in cells of mineralized tissue, facilitates mineralization, and is an indicator of bone metabolism.^{[ 76 ]} OPG is a decoy receptor in the RANKL/RANK pathway inhibiting osteoclast formation and thus limiting bone resorption.^{[ 77 ]} We observed differences in normalized ALP activity and OPG secretion between passage schemes suggesting the impact of the passage scheme on downstream osteoclast activity and bone resorption (Figure 4A,B).^{[ 77} , ^{78 ]} We further evaluated mineral deposition at Day 56 and observed cells from female donors deposit higher amounts of minerals than those from male donors in the p4‐5 group (Figure 4C,D). Mineral content at Day 56 was similar between passage schemes (Figure S3C, Supporting Information), and this suggests that cells are influencing mineral content similarly between passage schemes despite differences in normalized metabolic activity and cell number.

Next, we quantified osteogenic gene expression via a custom NanoString code set, where we observed cells from male donors displaying significantly higher expression levels at Day 56 compared to cells from female donors (Figure 5A; Figures S4A,B, Supporting Information). Most donors had similar osteogenic gene expression levels between passage schemes (Figure S4A, Supporting Information). This aligns with previous findings from our lab that the mineralized collagen scaffolds provide significant instructive cues to promote osteogenesis in vitro.^{[ 50} , ^{53 ]} Finally, we quantified secreted osteogenic factor release via a Luminex assay and observed cells from female donors secreted significantly more SPARC than male cells by Day 56 (Figure 6A). This suggests that this cohort of male and female cells have similar secretory potential of osteogenic factors. We have observed consistent osteogenic output on our scaffolds across species both in vitro^{[ 49} , ⁵⁰ , ⁷⁹ , ^{80 ]} and in vivo,^{[ 55} , ⁵⁷ , ^{59 ]} and these results further confirm the efficacy of our mineralized collagen scaffolds for bone tissue engineering. Regardless, it will be essential in future efforts to perform in vivo experiments to evaluate patterns of bone healing in the context of potential sex‐associated variability in MSC osteogenic capacity.

Additionally, we examined hMSC immunomodulatory and angiogenic potential via gene expression and soluble factors released. Cells from male donors had significantly higher immunomodulatory gene expression compared to female cells in both passage schemes at Day 56 (Figure 5B; Figure S4C, Supporting Information). Furthermore, cells from male donors secreted significantly higher amounts of immunomodulatory cytokines (four out of six analyzed cytokines) compared to female cells (Figure 6B) consistent with our gene expression data. This suggests there are significant sex‐associated variations in the immunomodulatory potential of these hMSCs, and this has been observed in vitro^{[ 81} , ^{82 ]} by other groups. As seen in Table 1, some donors had co‐morbidities while for others such health information was not provided. Thus, these differences in immunomodulatory potential could be attributed to different treatments, stress levels, and lifestyle habits that these patients experienced. In our previous work, we observed that stimulating hMSCs with inflammatory cytokines significantly enhances their immunomodulatory potential compared to a basal control^{[ 83 ]} therefore, we would expect variability in hMSC immunomodulatory potential as a function of the systemic inflammatory state in patients. VEGFA gene expression was higher in cells from male donors in the p4‐5 passage scheme and higher in female cells in the p3‐5 passage scheme with minimal variability in expression between passage schemes (Figure 5C, Figure S4E,F, Supporting Information). However, we saw no significant sex differences in VEGF secretion in p3‐5 (Figure 6C). Given the limited number of angiogenic genes and growth factors characterized here, it is difficult to make broad conclusions, motivating future expanded studies to characterize the impact of donor and sex on angiogenic potential. Differences in gene expression patterns between male and female donors were observed at a higher frequency in passage scheme p4‐5, suggesting the importance of passage length on cellular outputs, but most of the significant differences as a function of donor sex were consistent between passage schemes. Together, these results suggest donor and sex‐associated variations in cell activity and osteogenic output are significant and important to consider.

Overall, we observe that over the course of 56 days, the passage scheme has little influence on osteogenic potential (ALP activity, mineral content, and gene expression) but we do observe critical differences in donor‐to‐donor as well as evidence of sex‐associated variation in osteogenic potential. Altogether, these data suggest that cells from female donors possess a higher (p4‐5) or equivalent (p3‐5) osteogenic capacity in mineralized collagen scaffolds compared to male cells despite displaying lower proliferation than male cells (p4‐5 and p3‐5). Thus, there may be a difference between male and female patient needs when it comes to therapeutic intervention, where male patients might require a greater boost in osteogenic activity while female patients might require an enhancement in their cell proliferation. A limitation of this study, and an important consideration when studying variability as a function of sex, is the consideration of sex hormones. Estrogen is known to play vital roles in inhibiting bone resorption^{[ 84 ]} but has also been shown to interact with the immune system.^{[ 85 ]} Although in this work we did not directly study the role of estrogens and other sex hormones, we conducted these experiments in phenol‐red‐free media because phenol‐red is known to be a weak estrogen that can stimulate estrogen‐sensitive cells.^{[ 86 ]} However, future studies should expand these efforts and study sex‐based differences in hormone‐stripped media as serum additives provide significant hormone levels to the system. A second major concern is that our identification of sex variations is predicated on the reported sex of each donor. Sex exists beyond a simple binary, so it is imperative that bioengineering efforts continue evolving to carefully consider potential variability as a function of sex.

These 56‐day experiments are time‐consuming, large, and costly, and not all research labs have the resources needed to conduct such large experiments. The scale and cost of these experiments are a primary deterrent in incorporating multiple donors in an experiment and considering sex as an experimental variable. Ongoing efforts in our lab are focused on developing high‐throughput platforms to allow for the rapid evaluation of donor variability and donor characteristics as biological variables in 3D mineralized collagen biomaterials. Finally, sourcing enough cells from a large pool of donors is an additional challenge in conducting these experiments. Thus, this work is a call to action for academics and industry to expand their donor pools across sex, age, race, and pre‐existing conditions and make these cells more accessible to researchers. Access to diverse pools of donor cells would allow for the broadening of knowledge and reporting of findings considering subpopulations of people without generalizations that could lead to more harm.

4. Conclusion

Human mesenchymal stem cell (hMSC) passage number and culture time post‐thawing are important variables to determine hMSC osteogenic response, and it is essential that these are reported accurately. To this end, we evaluated differences in hMSC response as a function of passage number and culture length. Our findings show that the passage scheme can markedly affect hMSC response suggesting it is essential that culture parameters are reported in the literature. Another important factor when designing biomaterials for bone repair is donor‐ and sex‐associated variability. We want to ensure our mineralized collagen scaffolds provide an appropriate environment for patients across many axes of variation (age, sex, race/ethnicity, and health status). To this end, we evaluated the variability in osteogenic activity between 8 donors cultured on mineralized collagen scaffolds. An osteogenic response was observed for all donors on our mineralized collagen scaffolds in the absence of any exogenous osteogenic stimulant (BMP2, osteogenic media), showing their applicability to a large pool of potential patients. We report significantly increased cell proliferation and osteogenic gene expression from male donors and significantly higher mineral deposition from female donors. Further, we observe the frequency of sex differences is dependent on the passage scheme. These findings suggest mineralized collagen scaffolds provide an important in vitro platform to study donor variability and donor characteristics such as donor‐reported sex as experimental variables. Further, our work illuminates the need to expand the literature by studying donor‐ and sex‐associated variability and reporting culture methods to allow for transparency of data and results.

5. Experimental Section

Mineralized Collagen Scaffold Fabrication

Mineralized collagen‐glycosaminoglycan scaffolds were fabricated via lyophilization from a mineralized collagen precursor suspension. The mineralized collagen suspension was created by homogenizing type I collagen (Sigma Aldrich, St. Louis, Missouri USA), chondroitin‐6‐sulfate (Sigma Aldrich), and calcium salts (calcium hydroxide and calcium nitrate, Sigma Aldrich) in mineral buffer solution (0.1456 m phosphoric acid/0.037 m calcium hydroxide). The precursor suspension was stored at 4 °C and degassed prior to lyophilization.

Scaffolds were subsequently fabricated via lyophilization using a Genesis freeze‐dryer (VirTis, Gardener, New York USA).^{[ 87 ]} Briefly, 67.4 mL of precursor suspension was pipetted into a custom stainless‐steel mold (5 inches ×5 inches). Mineralized collagen sheets were then fabricated by freezing the suspension via cooling from 20 to −10 °C at a constant rate of 1 °C min⁻¹ followed by a temperature hold at −10 °C for 2 h. The frozen suspension was then sublimated at 0 °C and 0.2 Torr, resulting in a porous scaffold network. After fabrication, a 6 mm biopsy punch was used to obtain scaffolds that would be used for cell culture (6 mm diameter and 3 mm height). All scaffolds used for cell culture were sterilized via ethylene oxide treatment for 12 h utilizing an AN74i Anprolene gas sterilizer (Andersen Sterilizers Inc., Haw River, North Carolina USA) in sterilization pouches.^{[ 44} , ⁵⁰ , ^{88 ]} All subsequent handling steps leading to studies of cell activity were performed in a sterile manner.

Mineralized Collagen Scaffold Hydration

Dry sterile scaffolds were hydrated and crosslinked prior to cell seeding as previously described^{[ 42} , ⁶¹ , ^{79 ]} and all steps were done under moderate shaking. Scaffolds were first hydrated for 2 h in ethanol at ≈20 °C (room temperature). Scaffolds were then washed in PBS for 1 h at ≈20 °C. Next, scaffolds were crosslinked for 2 h in EDC‐NHS at ≈20 °C and washed again in PBS. Finally, the scaffolds were incubated in a cell culture medium for 48 h at 37 °C with a complete media change after 24 h.

Human Mesenchymal Stem Cell Culture—Human Mesenchymal Stem Cell (hMSC) Source

Human mesenchymal stem cells (hMSCs) were purchased from Essent Biologics (Centennial, Colorado USA) and RoosterBio (Frederick, Maryland, USA). Cells from Essent Biologics were derived from cadaveric adipose tissue and cells from RooserBio were derived from bone marrow donors. Table 1 contains donor ID, donor‐reported sex, age, ethnicity, pre‐existing conditions, tissue origin, company, and lot number for each cell used. Passage 2 hMSCs were expanded in T175 flasks (Thermo Fisher Scientific, Hampton, New Hampshire USA) and were cultured until passage 3 and 4 in RoosterNourishTM‐MSC expansion medium (RoosterBio) at 37 °C and 5% CO2. Once at passage 3 and 4 cells were frozen and stored in liquid nitrogen until further use.

Human Mesenchymal Stem Cell Culture—Culturing hMSCs to Compare Passage Number and Culture Time Influences on Cell Response

Passage 3 hMSCs were randomly chosen to obtain two male and two female donors (donor‐reported sex). Cells were thawed and cultured in T175 flasks (Thermo Fisher Scientific, Hampton, New Hampshire USA) supplemented with RoosterNourishTM‐MSC expansion medium (RoosterBio) at 37 °C and 5% CO2. These passage 3 (p3) cells were expanded for two passages to p5 or for three passages to p6. These will be referenced as p3‐5 and p3‐6 in this manuscript. Passage 4 hMSCs were thawed and cultured in T175 flasks supplemented with RoosterNourish‐MSC expansion medium for one passage to p5. These cells with be referenced as p4‐5 in this manuscript. Once confluent, p3‐5, p3‐6, and p4‐5 hMSCs were seeded (150 000 cells/scaffold) onto mineralized collagen scaffolds using an orbital seeding method^{[ 89} , ^{90 ]} for 6 hours at 37 °C and 5% CO2. Cell‐seeded scaffolds were transferred to ultra‐low attachment plates and cultured in mesenchymal stem cell growth media (phenol red‐free low glucose DMEM + L‐glutamine, 10% fetal bovine serum, 1% antibiotic‐antimycotic) at 37 °C and 5% CO2 for 21 days (Figure 1A). Every 3 days, the media was completely changed and 300 uL of media was stored at −20 °C for further analysis (ELISA).

Human Mesenchymal Stem Cell Culture—Culturing hMSCs to Compare Donor–Donor Variability

Passage 3 and passage 4 cells from eight donors (four male and four female) were expanded in T175 flasks (Thermo Fisher Scientific, Hampton, New Hampshire USA) and were cultured until passage 5 (p3‐5 and p4‐5, respectively) in RoosterNourish‐MSC expansion medium (RoosterBio) at 37 °C and 5% CO2. Once confluent, p3‐5 and p4‐5 hMSCs were seeded (150 000 cells/scaffold) onto mineralized collagen scaffolds using an orbital seeding method^{[ 89 ]} for 6 hours at 37 °C and 5% CO2. Cell‐seeded scaffolds were transferred to ultra‐low attachment plates and cultured in mesenchymal stem cell growth media (phenol red‐free low glucose DMEM + L‐glutamine, 10% fetal bovine serum, 1% antibiotic‐antimycotic) at 37 °C and 5% CO2 for 56 days (Figure 1B). Every three days, the media was completely changed and 300 uL of media was stored at −20 °C for further analyses (ELISA, Luminex, ALP activity).

Cell Number and Cell Metabolic Activity

Cell numbers were obtained via DNA isolation on Days 7, 14, and 21. DNA extraction was completed using a DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany). Cell‐seeded scaffolds were cut into quarters and placed into 360 uL Buffer ATL and 40 uL proteinase K and incubated at 56 °C for 18–24 h. Then, 400 uL Buffer AL and 400 uL ethanol were added to the lysates and vortexed. The remaining steps were followed per the manufacturer's instructions. DNA concentration was quantified using a NanoDrop Lite Spectrophotometer (Thermo Fisher Scientific) and cell number was determined from standard curves generated with known cell numbers for p3‐5, p3‐6, and p4‐5.

Metabolic activity was obtained using an alamarBlue viability assay (Invitrogen, Carlsbad, California USA) on Days 7, 14, and 21. Cell‐seeded scaffolds were incubated in a 10% alamarBlue solution for 90 min at 37 °C under moderate shaking. The fluorescence of the alamarBlue solution was then read using a fluorescent spectrophotometer (Tecan Infinite F200 Pro, Männedorf, Switzerland). The active ingredient in the alamarBlue reagent, resazurin, is reduced to a compound that is highly fluorescent when added to metabolically active cells.^{[ 50} , ^{91 ]} The relative cell metabolic activity was determined from standard curves generated with known hMSC concentrations for p3‐5, p3‐6, and p4‐5. The relative cell metabolic activity was reported as a fraction of the initial seeding cell count (i.e., an experimental value of one indicates the metabolic activity of the number of cells seeded onto the scaffold) and normalized by cell number.

RNA Isolation from Mineralized Collagen Scaffolds

RNA was isolated from cells at Day 0 using the RNAqueous Total RNA Isolation Kit (Invitrogen), and the RNA was eluted in 40 uL Elution Solution. RNA was isolated from cell‐seeded scaffolds using TRIzol (Thermo Fisher Scientific) and the RNeasy Mini Kit (Qiagen). Cell‐seeded scaffolds were cut into quarters and placed into Phasemaker Tubes (Invitrogen). Next, 1 mL TRIzol was added to each tube, and scaffolds with TRIzol were vortexed and incubated at room temperature for 5 min. Then, 200 uL chloroform was added to each tube, and scaffolds with TRIzol and chloroform were vortexed and incubated at room temperature for 3 minutes. Tubes were vortexed immediately before centrifuging at 15 000 g for 15 min at 4 ⁰C. After centrifugation, the aqueous phase was added to 650 uL 70% ethanol and mixed briefly. This solution was then pipetted into the RNeasy Mini Kit extraction columns and the kit instructions were followed as per the manufacturers recommendation. RNA was eluted in RNAse‐free water and quantified using a NanoDrop Lite Spectrophotometer (Thermo Fisher Scientific). RNA samples were stored at −80⁰C until further analysis.

Nanostring Gene Expression Evaluation

Transcript expression was quantified with the NanoString nCounter System (NanoString Technologies, Inc.) located at the Tumor Engineering and Phenotyping Shared Resource at the Cancer Center in Illinois using a custom panel of 38 mRNA probes (Table S1, Supporting Information). The NanoString nCounter System identifies and counts individual transcripts using unique color‐coded probes, and there were no reverse transcription or amplification steps required. Isolated RNA was quantified and quality tested using Qubit RNA BR Assay Kit, loaded into cartridges, and run on the NanoString assay as instructed by the manufacturer. The nSolver Analysis Software (NanoString Technologies, Inc.) was used for data processing, normalization, and evaluation of expression. Raw data was normalized to Day 0 controls (n = 3). Expression levels are depicted as log twofold change.

Enzyme‐Linked Immunosorbent and Luminex Assays

Media was collected every three days and pooled at Day 3, Day 9 (days 6–9), Day 15 (days 12–15), Day 30 (days 18–30), and Days 56 (days 33–56). An enzyme‐linked immunosorbent assay (ELISA) (R&D Systems, Minneapolis, MN) was used to quantify the amount of osteoprotegerin (OPG) released by cells seeded on mineralized collagen scaffolds (n = 6). The ELISA was conducted based on the manufacturer's protocol and the absorbance readings were done with a spectrophotometer (M200, Tecan, Switzerland). Concentrations were converted from absorbance values using a standard curve. A custom Luminex panel (R&D Systems) was used to quantify the amount of 14 proteins released by cells seeded on mineralized collagen scaffolds (n = 4) (Table S2, Supporting Information). The LUMINEX was conducted as per manufacturer protocol, absorbance was read on a Luminex 200 XMAP system (Luminex Corporation, Austin, TX), and each analyte was normalized to a background control and converted to a concentration using a standard curve for each factor. Expression levels of all secreted proteins were depicted as cumulative concentrations.

Alkaline Phosphatase (ALP) Assay

Media was collected every three days and pooled at Day 3, Day 9 (days 6–9), Day 15 (days 12–15), Day 30 (days 18–30), and Days 56 (days 33–56). An ALP activity assay (Abcam, England) was used to determine cell‐dependent osteogenic activity (n = 6). Results were compared between media isolated from cell‐seeded scaffolds for the entire length of culture, using unseeded scaffolds as a background control. P‐nitrophenyl phosphate (pNPP, µmol) concentration per well was converted to U mL⁻¹ with known reaction time and volume of sample and then normalized by cell number.

Quantification of Calcium and Phosphorus in Freeze‐Dried Scaffolds

Scaffolds for each donor were harvested on Day 56 (n = 6), washed in PBS, fixed in Formal‐Fixx (ThermoFisher Scientific, Waltham, MA, USA) at 4 °C overnight, and washed in PBS three times for 5 min at room temperature under moderate shaking. The scaffolds were then frozen at −80 °C overnight and freeze‐dried using a Genesis freeze‐dryer (VirTis, Gardener, New York USA). Freeze‐dried scaffolds were analyzed via inductively coupled plasma mass spectrometry (ICP‐MS). Dry samples were transferred to a digestion tube, then digested with concentrated nitric acid (67–70%) followed by automated sequential microwave digestion in a CEM Mars 6 microwave digester (CEM Microwave Technology Ltd., North Carolina, USA). The final product was a clear aqueous digest which was diluted to a volume of 25 mL using DI water. This solution was then introduced to an inductively coupled plasma‐mass spectrometer (NexION 350D ICP‐MS, PerkinElmer, USA) for the elemental analysis in a standard mode.

Statistics

RStudio was used for all plotting (ggplot2) and statistical analysis. No outliers were removed. The sample size was six (n = 6) for alamarBlue, ELISA, ALP, and ICP‐OES, four (n = 4) for Luminex, and three (n = 3) for NanoString and cell number. All experiments had three or more experimental groups. Residuals were tested for normality using the Shapiro–Wilk test, and homogeneity of variance was tested using the Levene Test.

For passage scheme experiments (Figure 2): when both assumptions were met, we ran a one‐way ANOVA (independent variable: passage scheme) and Tukey's HSD mean separation with alpha = 0.05, when data did not have equal variance, a one‐way ANOVA was run with Welch's correction and Tukey's HSD mean separation with alpha = 0.05, when data were not normally distributed we ran a one‐way ANOVA and Tukey's HSD mean separation with alpha = 0.01, and when both assumptions were unmet we ran Mood's median for nonparametric data and median separation with alpha = 0.05. For donor variability and sex influence experiments (Figures 3, 4, 5, 6; Figures S1–S4, Supporting Information): when both assumptions were met, we ran a two‐way ANOVA (independent variables: donor and sex), and Tukey's HSD mean separation with alpha = 0.05, and if one or both assumptions were unmet we ran a two‐way ANOVA and Tukey's HSD mean separation with alpha = 0.01.

Conflict of Interest

The authors declare no conflicts of interest.

Author Contributions

V.K. and A.T. contributed equally to this work. V.K. performed conceptualization, data curation, formal analysis, visualization, investigation, methodology, wrote original draft, wrote reviewed and edited. A.T. performed conceptualization, data curation, formal analysis, visualization, investigation, methodology, wrote original draft, wrote reviewed and edited. M.P. performed investigation, data curation, formal analysis. B.H. performed conceptualization, resources, project administration, funding acquisition, supervision, wrote reviewed and edited.

Supporting information

Supporting Information

Kolliopoulos V., Tiffany A., Polanek M., Harley B. A. C., Donor Sex and Passage Conditions Influence MSC Osteogenic Response in Mineralized Collagen Scaffolds. Adv. Healthcare Mater. 2024, 13, 2400039. 10.1002/adhm.202400039

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.