Introduction

Large bone defects caused by trauma or therapeutic resection are difficult to treat. Often, the defect void is filled with bone graft to provide structural integrity, osteo-conduction, osteo-induction, or a combination thereof. Autograft bone harvested from the patient yields the best clinical outcomes. However, the amount of autograft that can be harvested is limited, and donor site morbidity is a significant clinical concern. Alternative approaches may include using the Masquelet method to generate additional tissue for grafting or employing limb-lengthening procedures to avoid grafting altogether [1–4]. Allograft and synthetic bone graft materials are commonly used to treat large bone defects [5]. Allograft is generally abundant and available in different physical forms that can be tailored to the required need. However, the ability of allograft or synthetic bone graft to promote osteogenesis within a bone defect is significantly less than that of autograft [6, 7].

Zinc is an essential nutrient that concentrates in bone and appears to have one or more physiological roles in the skeleton [8]. Previous studies found that dietary Zn deficiency leads to reduced levels of Zn in bone and can impair adolescent growth in humans and animal models [9–13]. Clinical studies have found that elevated serum Zn levels correlate with increased bone mineral density and reduced fracture risk [14]. Conversely, patients with osteoporosis were found to have reduced serum Zn levels and reduced bone tissue levels of Zn [15, 16]. Animal models and in vitro studies have also found that Zn can inhibit osteoclast activity and stimulate osteoblast activity [17–20].

Zinc can also promote bone regeneration when applied to sites of bone healing. Using a rat femur fracture model, local application of Zn to the fracture site significantly improved healing as shown by enhanced callus cellularity, increased callus levels of insulin-like growth factor-1 and vascular endothelial growth factor, increased percentages of callus cartilage and bone, and increased callus mechanical properties, demonstrating the potential therapeutic use of Zn for bone regeneration [21–23]. Local application of Zn with autograft improved posterolateral spinal fusion in a rat model [24]. Several studies have also noted that incorporating Zn into synthetic materials can enhance the osteogenic activity of those materials [25–27]. A recent experiment also found that adsorbing Zn to morselized, processed human allograft significantly improved osteogenesis in critically-sized rat femoral defects as compared to non-adsorbed control allograft [28].

Zinc also has anti-microbial properties. Using zinc acetate, Atmaca et al. [29] found that soluble Zn inhibited Staphylococcus aureus (S. aureus) and Staphylococcus epidermidis (S. epidermidis) growth at concentrations above 2.2 mM but had no effect on the growth of Pseudomonas aeruginosa (P. aeruginosa). Gugala et al. [30] also found that approximately 2 mM Zn could inhibit the growth or biofilm formation of S. aureus, but was less effective against P. aeruginosa or Escherichia coli (E. coli). Inclusion of Zn into bioactive glass at 1.2% Zn by weight or hydroxyapatite at 1.6% Zn by weight inhibited the growth of Streptococcus mutans (S. mutans) and S. aureus, respectively [31, 32]. Thus, treating bone allograft with Zn may also provide a beneficial anti-microbial effect during and after allograft preparation.

The improved osteogenic capacity of the Zn-treated allograft suggests that binding of Zn to the allograft can be adjusted to improve clinical performance. To that end, this study measured the effects of Zn concentration, incubation time, potentially competitive divalent cations, and different Zn salts on Zn-binding to human allograft. By examining adsorption parameters that affect Zn binding to allograft, we aim to better understand how allograft Zn treatment can be optimized for future in vivo testing. The potential use of Zn-binding to allograft to reduce bacterial burden was also investigated.

Materials and methods

Processed human cortical cancellous bone allograft was obtained from a commercial vendor (SpineFrontier, Beverly, MA, USA). Tissue banks must obtain consent from the donor or the donor’s family before collecting any tissue. The allograft was pulverized and sieved to obtain 0.5 to 1 mm size granules, which were used in all subsequent experiments. Micro-spin filters [0.2 µm pore size polyvinylidene fluoride (PVDF) membrane, 850 µL capacity, polypropylene housing] in 2 mL microcentrifuge tubes were purchased from Analytical Sales and Services, Inc. (Flanders, NJ, USA). PVDF midi-spin filters (0.2 µm pore size membrane, 4 mL capacity, polypropylene housing) in 7 mL tubes were purchased from Ciro Manufacturing Corp. (catalog number BPV02, Deerfield Beach, FL, USA). Anhydrous zinc chloride and zinc acetate dihydrate were purchased from Sigma-Aldrich (St. Louis, MO, USA). Zinc picolinate was purchased from Spectrum Chemical Mfg. Corp. (New Brunswick, NJ, USA). The different Zn salts were dissolved in Tris-buffered saline (TBS) that consisted of 50 mM Tris-Cl and 150 mM NaCl, pH 7.4.

Staphylococcus aureus, subspecies aureus, strain NCTC 8532 (S. aureus) was purchased from the American Type Culture Collection (Manassas, VA, USA). S. aureus was cultured at 37°C in Difco Nutrient Broth media or agar plates (Fisher Scientific, Pittsburgh, PA, USA). Colonies of S. aureus from agar plates were used to inoculate 3 mL of Difco Nutrient Broth, that were cultured overnight at 37°C with shaking. The overnight cultures were first diluted to an optical density (OD₆₀₀) of 0.1 [equal to 10⁷ colony forming unit (CFU) per mL] in fresh Difco Nutrient Broth and then to 10³ CFU/mL in TBS in preparation for treating bone allograft. Only fresh, overnight cultures were used for experiments.

Results

Time and concentration effects on Zn-allograft binding

In a prior study, we found that treating morselized human bone allograft with 0.1% w/v ZnCl₂ for 30 min improved the osteogenic capacity of the allograft in a rat femur defect model [28]. Within that study, experiments showed that Zn binding to allograft occurred in a time- and concentration-dependent manner [28]. However, time-dependent binding was only measured using a relatively high concentration of Zn (1% w/v ZnCl₂), while concentration-dependent binding was only measured after an extended (24-hour) binding time.

To better characterize Zn-allograft binding, we measured Zn binding to morselized human bone allograft at lower ZnCl₂ concentrations (0.01%, 0.1%, and 1% w/v) and at shorter binding times (0.5, 3, and 24 hours; Figure 1A). Again, we found that Zn binding was concentration- (p < 0.001) and time-dependent (p < 0.001) between 0.01% and 1% w/v ZnCl₂ and 0.5 and 24 hours of binding time.

Display full size

Figure 1. Zn binding to bone allograft: [Zn] and time effects. Morselized human bone allograft was treated with ZnCl₂ or TBS. After washing, Zn was extracted from the allograft with nitric acid and measured by ICP-MS. (A) Allograft was treated with ZnCl₂ at 0.01%, 0.1%, and 1% (w/v) for 0.5, 3, or 24 hours or with TBS for 24 hours. (B) Allograft was treated with 0.1% or 0.5% w/v ZnCl₂ for 1, 3, 7, or 15 mins or with TBS for 15 mins. (C) Allograft was treated for 7 mins with TBS or 0.1, 0.3, 0.5, and 1% (w/v) ZnCl₂. (A, B) Values shown are the means (± SD) from 3 independent replicates (2 independent replicates for TBS and 6 independent replicates for 1% ZnCl₂ at 24 hours, panel A). (C) Bars represent the mean and SD of the 3 independent replicates for each Zn concentration. The amount of Zn bound per gram of allograft was compared to binding time (A, B) and to Zn concentration (A, B, C). p-values from the 2-way ANOVA analyses for effects of Zn concentration [Zn] and time on Zn bound to allograft are shown in each graph (A, B). p-values from a 1-way ANOVA for effects of Zn concentration [Zn] on Zn bound to allograft are shown (C)

Closer examination of the binding data in Figure 1A indicated that Zn binding occurred rapidly within the first 30 minutes, with respective averages of 1.2, 2.3, and 7.0 mg of Zn bound per g of allograft for the 0.01%, 0.1%, and 1% ZnCl₂ concentrations. Consequently, the potential effects of Zn concentration over shorter binding times were evaluated (Figure 1B). However, no effect of binding time was detected between 1, 3, 7, and 15 min of Zn binding to allograft bone when using ZnCl₂ concentrations (w/v) of 0.1% and 0.5% (p = 0.14; Figure 1B), nor was any effect detected when Zn concentrations (w/v) of 0.1%, 0.3%, 0.5%, and 1% ZnCl₂ were bound to allograft bone for 7 min (p = 0.60; Figure 1C).

Zn salt anion effects on bone allograft binding

As a divalent cation, Zn can form a variety of salts with differing water solubility (Table 1). The ability of different Zn salts to affect Zn adsorption to allograft was measured. Approximately 15 mM solutions of Zn, as Zn chloride (0.2% w/v), Zn acetate (0.32% w/v), and Zn picolinate (0.46% w/v) were used to treat morselized human bone allograft for 7, 20, or 60 min. As shown in Figure 2, the amount of Zn bound to the allograft increased with binding time for all salts (p < 0.001). The Zn salt used also affected the amount of Zn bound to the allograft (p < 0.001). Interestingly, the amount of Zn bound to the allograft was significantly greater after 7 min incubation in Zn acetate (0.82 mg of Zn per g of allograft) as compared to ZnCl₂ (0.44 mg of Zn per g of allograft). In contrast, the amount of Zn bound to allograft bone from Zn picolinate (0.25 mg of Zn per g of allograft) was significantly lower than that from Zn chloride after 7 min and from Zn acetate after 7, 20, and 60 min of binding.

Display full size

Figure 2. The effects of different Zn salts on Zn binding to bone allograft. Binding of Zn to allograft is dependent on the Zn salt used. Aliquots of morselized human allograft were incubated in approximately 15 mM solutions of zinc chloride, zinc acetate, or zinc picolinate for 7, 20, or 60 min at room temperature (n = 3 per salt and per time). Bound Zn was extracted with nitric acid and measured using ICP-MS. Shown are mean values (± SD). Overall, more Zn was bound to the allograft using zinc acetate (green triangles) after 7 min than from either zinc chloride (red circles) or zinc picolinate (blue inverted triangle). Conversely, less Zn was bound to allograft using zinc picolinate than from either zinc acetate or zinc chloride and at all time points tested. Statistical significance is indicated in the graph

Effect of MgCl₂ on Zn binding to bone allograft

The potential effect of other cations competing for Zn binding sites on the bone allograft was investigated using increasing concentrations of Mg (0–2% w/v of MgCl₂-6H₂O) and a constant concentration of 0.5% w/v ZnCl₂ (Figure 3). Only a single binding time of 7 min was investigated. The baseline level of Mg in the bone allograft (0.71 ± 0.15 mg/g) was approximately 4-fold higher than the average baseline level of Zn detected in bone allograft incubated in TBS (0.16 ± 0.12 mg/g, Figure 1). No difference in binding of Mg to the allograft was detected after 7 min of incubation for all Mg concentrations tested (ANOVA p = 0.064; linear regression p = 0.142, Figure 3A). However, a significant difference was detected when comparing allograft samples incubated with 0% Mg (0.71 ± 0.15 mg/g) to all allograft samples treated with ≥ 0.1% Mg (0.96 ± 0.10 mg/g; t-test, p = 0.003), indicating that approximately 0.25 mg of Mg bound per g of bone allograft.

Display full size

Figure 3. Divalent cation effects on Zn binding to bone allograft. Rapid binding of Zn to bone allograft is not affected by Mg concentration. Aliquots of bone allograft were treated for 7 min at room temperature with 0.5% (w/v) ZnCl₂ in TBS and containing 0%, 0.1%, 0.5%, 1%, or 2% (w/v) MgCl₂-6H₂0. Mg (panel A) and Zn (panel B) that remained bound to the allograft were extracted using nitric acid and measured by ICP-MS. Shown are mean values (± SD) from 3 independent replicates for each Mg concentration. Linear regression found no significant effect of increasing Mg levels on Zn binding after 7 min of binding time (B, p = 0.066)

The addition of Mg to the Zn solution did not inhibit Zn binding to the bone allograft during the 7 min of binding (Figure 3B). Comparing the amount of Zn bound from 0.5% w/v ZnCl₂ with no added MgCl₂ (0.495 ± 0.079 mg/g) to the amount of Zn bound from the 0.5% ZnCl₂ with ≥ 0.1% w/v MgCl₂ (0.655 ± 0.068 mg/g) found that in the presence of MgCl₂, 32% more Zn bound to the allograft during the 7 min binding reactions (t-test, p = 0.004).

Effect of pH on Zn binding to bone allograft

The potential effect of buffer pH on Zn binding to bone allograft was measured. The pH of TBS was adjusted to yield solutions in the pH range of 6 to 8. ZnCl₂ was dissolved in the pH adjusted TBS to 0.5% (w/v). No effect on pH after dissolving the ZnCl₂ was detected. Aliquots of bone allograft were treated with pH adjusted TBS-ZnCl₂ solutions for 30 min at room temperature with shaking. The amount of Zn bound to the allograft was measured (Figure 4). The amount of Zn bound remained constant over the pH range tested (p = 0.42).

Display full size

Figure 4. pH-independent Zn binding to bone allograft. Aliquots of morselized human allograft were treated with 0.5% (w/v) ZnCl₂ dissolved in TBS buffered to pH 6.0, 6.3, 6.6, 7.0, 7.4, or 8.0 for 30 min and then washed extensively with TBS buffered to the same pH. Zn that remained bound to the allograft was extracted with nitric acid and measured by ICP-MS. Shown are mean (± SD) values from 3 independent replicates for each pH tested. A comparison of the amount of Zn bound to allograft found no statistical difference between the groups (p = 0.42)

Zn binds to the surface of bone allograft

The rapid binding of Zn to the bone allograft indicated that binding was occurring primarily on the surface of the bone allograft morsels. To confirm that Zn was bound to the surface, allograft was bound with 0.5% ZnCl₂ for 1 hour. The bound Zn was then detected using Zinpyr-1, which fluoresces green when bound to Zn [33]. As shown in Figure 5, Zinpyr-1 fluorescence was primarily detected on the surface of the allograft morsels, indicating that Zn binding occurred on the surface of the allograft morsels.

Display full size

Figure 5. Surface binding of Zn to bone allograft. Morselized human allograft was treated with 0.5% (w/v) ZnCl₂ dissolved in TBS buffered to pH 7.4 for 60 min and then washed extensively with TBS before detecting bound Zn with the fluorescent stain, Zinpyr-1. The stained allograft morsels were polished to reveal the interior of the bone morsels before epi-fluorescent detection of the Zinpyr-1. Shown are two, approximately 1 mm diameter allograft morsels; scale bars are 200 µm

Zn-bound allograft is resistant to Staphylococcus aureus

Binding Zn to allograft bone is likely to alter the surface chemistry of the allograft. The bound Zn could potentially reduce bacterial adherence or present a sufficiently high local Zn concentration to inhibit bacterial growth. To test this, morselized human bone allograft was incubated in 0.5% ZnCl₂ or TBS for 1, 7, 15, or 60 min, washed multiple times with TBS pH 7.4, dried, and then sterilized by gamma-irradiation. The allograft aliquots were inoculated with S. aureus and, after washing, cultured at 37°C with shaking in Nutrient Broth. The optical density (OD₆₀₀) of each allograft-S. aureus culture was measured from 9 to 14 hours to determine the growth of S. aureus and, therefore, the relative amount of viable S. aureus retained by the allograft. Growth curves for S. aureus were plotted for each allograft culture and the AUC was calculated for each growth curve (Figure 6). Growth curves appeared similar between control allograft and allograft treated for 1 min with ZnCl₂, but apparent reductions in S. aureus growth were evident in cultures of allograft treated with ZnCl₂ for 7 min or more (Figure 6A–D). AUCs were significantly different between control and Zn-treated allografts for all ZnCl₂ treatment times tested (Figure 6A–D). A comparison of all the control allograft AUCs and all the Zn-treated allograft AUCs showed a significant reduction in S. aureus growth in the Zn-treated allograft cultures (Figure 6E).

Display full size

Figure 6. Staphylococcus aureus growth curves in bone allograft cultures. Bone allograft samples were incubated in 0.5% (w/v) ZnCl₂ or TBS for 1, 7, 15, or 60 minutes, washed, and then inoculated with Staphylococcus aureus. After 30 min, the allograft samples were extensively washed with TBS before bacterial growth media was added. Bacterial growth was measured by optical density (OD₆₀₀) from 9 to 14 hours after the addition of bacterial media. Panels A–D show the mean (± SD) OD₆₀₀ values from 9 to 14 hours after Nutrient Broth addition for the 1-, 7-, 15-, and 60-min treatment times (respectively, panels A–D, left graph). AUCs were calculated from 9–14 hours for each allograft aliquot. Within each treatment time, the AUCs were compared (panels A–D, right graph). All treatment times with ZnCl₂ showed a reduction in Staphylococcus aureus growth based on the AUCs (p values are indicated in each graph). Panel E shows the AUCs from all control- and Zn-treated bone allograft cultures. An unpaired t-test was used to compare the AUC values for the control-treated and Zn-treated allograft cultures, which detected a significant reduction in Staphylococcus aureus growth (p < 0.0001). Green triangles represent the control, TBS-treated allograft cultures. Red, inverted triangles represent the Zn-treated allograft cultures

Discussion

The experimental results indicate that the amount of Zn binding to the allograft is not dependent on ZnCl₂ concentration (from 0.1% to 1% w/v) when treatment time is relatively short (< 15 min, Figure 1). However, Zn binding to allograft is dependent on ZnCl₂ concentration when treatment time is longer (1 hour or more) and the ZnCl₂ solution is > 0.1% ZnCl₂ (Figure 1) [28]. This suggests that Zn binds to allograft bone through more than one modality. As bone is composed of both organic and mineral components, the fast, concentration-independent binding and slow, concentration- and time-dependent binding of Zn may reflect differential Zn binding affinities between the organic and mineral components of bone or between surface binding of Zn to bone hydroxyapatite versus Zn substitution of calcium within the hydroxyapatite crystal [34]. In support of the former, a study from Fuierer et al. found that Zn bound to the surface of synthetic, non-sintered hydroxyapatite, most likely as a zinc phosphate surface precipitate, while Haumont detected Zn in the calcifying cartilage of rat growth plates and the osteoid (non-mineralized bone protein matrix) of canine Haversian canals [35, 36]. We found that even after 1 hour of incubation in 0.5% ZnCl₂, binding of Zn to the bone allograft occurred primarily on the surface (Figure 5). The presence of Tissue Non-specific Alkaline Phosphatase (TNSALP) at sites of calcifying cartilage and mineralizing osteoid would contribute to Zn localization at these sites since Zn is a co-factor for TNSALP [37, 38]. However, phosphate produced from TNSALP enzymatic activity at sites of calcifying cartilage and osteoid may amplify local Zn levels by precipitating Zn as zinc phosphate. Interestingly, Fuierer et al. [36] also found that Zn binding to the surface of hydroxyapatite inhibited further hydroxyapatite crystal growth. Thus, controlling soluble levels of Zn at sites of active bone formation may be a means to regulate bone mineral deposition.

Differences in Zn binding to allograft from the different zinc salts were somewhat unexpected (Table 1 and Figure 2). The expectation was that the aqueous solubility of each zinc salt would correlate with the concentration of Zn²⁺ ions available to bind to the bone allograft. As expected, the low aqueous solubility of zinc picolinate leads to reduced binding of Zn to the allograft bone as compared to the highly aqueous soluble zinc chloride. In contrast, the moderately soluble zinc acetate produced more Zn binding to the bone allograft than the highly soluble zinc chloride. The difference between the rapid binding of Zn to bone allograft from zinc acetate versus zinc chloride is not known. However, the ability of Zn²⁺ to interact with water appears to be affected by how well the anion can buffer in order to prevent Zn precipitation [39]. Since the buffering capacity of acetate is much greater than chloride, perhaps this could account for the faster binding of Zn to bone allograft from a zinc acetate solution than a zinc chloride solution.

The mechanism by which Zn binds to the allograft is not known. Spadaro et al. [40] found that Zn bound to both the organic (collagen) and mineral phases of bone, indicating multiple mechanisms for sequestering Zn in bone tissue. Several trace metals are found in normal bone tissue, and among those trace metals, Mg and Zn are two of the more abundant elements [41, 42]. Mg and Zn dissolve as divalent cations. Thus, if Zn binding to the bone allograft was strictly through an ionic interaction, excess levels of Mg²⁺ should compete for Zn²⁺ binding. However, we detected no competitive effect of Mg on Zn binding even at an approximate 2.5-fold excess of Mg (0.5% w/v ZnCl₂ and 2.0% w/v MgCl₂-6H₂O, Figure 3). In addition, binding of Zn was independent of pH over the pH 6 to pH 8 range tested (Figure 4). This indicates that Zn binding to the allograft has more specificity than a simple ionic interaction.

In addition to improving the osteogenic activity of processed bone allograft [28], Zn binding to allograft also appears to reduce the ability of bacteria to attach to allograft or to remain viable when attached to the Zn-bound allograft. As shown in Figure 6, Zn-bound allograft was resistant to binding viable S. aureus, one of the most common pathogens associated with osteomyelitis [43]. While even 1 min of treating the allograft with 0.5% w/v ZnCl₂ reduced viable S. aureus binding, longer ZnCl₂ incubation times appeared to be more effective.

Use of Zn in orthopaedic applications is an area of increasing interest because of the positive effects of Zn on osteogenesis and because Zn can be used to make biodegradable metal fixation devices [8, 44]. Use of Zn as an allograft adjuvant to enhance bone regeneration offers significant advantages over other approaches. Unlike growth factors or other protein adjuvants, Zn should not be affected by standard sterilization procedures and would allow for convenient, room temperature storage of the Zn bound allograft. As noted, Zn is an essential trace mineral that is necessary for multiple molecular processes. An estimated 10% of the human proteome binds Zn [45]. Sufficient amounts of Zn are generally consumed through a normal diet, but oral supplements are commonly used. The recommended daily allowance for Zn consumption in the United States is 8 mg/day for women and 11 mg/day for men, with a recommended maximum of 40 mg/day. The LD₅₀ for oral Zn toxicity is 3 g/kg body weight [46, 47]. Zn is used therapeutically for topical and enteric applications. Zn is considered a generally recognized as safe and effective substance by the US Food and Drug Administration. In our prior work, we found that Zn bound to allograft at 0.4% w/w was effective in promoting bone formation, which would equate to 40 mg of Zn per 10 g of bone allograft [28]. At the 0.4% w/w dose level of Zn to allograft and given the apparent slow-release of Zn from allograft, use of Zn-bound allograft to treat skeletal injuries would be unlikely to produce any safety concerns.

The results of this study will be useful for determining how best to restore trace metals or other substances to bone allograft that are lost during processing. In addition, the results of this study indicate that trace substances do not necessarily bind to bone through simple ionic interactions with the hydroxyapatite mineral phase of bone. Whether the binding mechanism utilized to restore Zn levels to processed allograft is similar to how Zn normally binds to bone in vivo will require additional investigation. Still, the results of this study can be used to help optimize methods that will enhance the osteogenic activity and bacterial resistance of bone allograft to provide better clinical outcomes.