Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Orthop Res. Author manuscript; available in PMC 2010 July 1.
Published in final edited form as:
PMCID: PMC2735874

Injection Temperature Significantly Affects In Vitro and In Vivo Performance of Collagen-Platelet Scaffolds


Collagen-platelet composites have recently been successfully used as scaffolds to stimulate anterior cruciate ligament (ACL) wound healing in large animal models. These materials are typically kept on ice until use to prevent premature gelation; however, with surgical use, placement of a cold solution then requires up to an hour while the solution comes to body temperature (at which point gelation occurs). Bringing the solution to a higher temperature before injection would likely decrease this intra-operative wait; however, the effects of this on composite performance are not known. The hypothesis tested here was that increasing the temperature of the gel at the time of injection would significantly decrease the time to gelation, but would not significantly alter the mechanical properties of the composite or its ability to support functional tissue repair. Primary outcome measures included the maximum elastic modulus (stiffness) of the composite in vitro and the in vivo yield load of an ACL transection treated with an injected collagen-platelet composite. In vitro findings were that injection temperatures over 30°C resulted in a faster visco-elastic transition; however, the warmed composites had a 50% decrease in their maximum elastic modulus. In vivo studies found that warming the gels prior to injection also resulted in a decrease in the yield load of the healing ACL at 14 weeks. These studies suggest that increasing injection temperature of collagen-platelet composites results in a decrease in performance of the composite in vitro and in the strength of the healing ligament in vivo and this technique should be used only with great caution.

Keywords: collagen, hydrogel, rheometry, ACL, tissue engineering

Recent studies have shown the efficacy of using collagen-platelet composites (CPCs) to stimulate anterior cruciate ligament (ACL) healing after partial and complete transection in animal models.13 These composites are thought to serve as a scaffold in the ACL wound site, filling the gap between the ruptured ends of the ACL, releasing growth factors, and promoting cellular proliferation and migration into the scaffold.2,3 The use of collagen-based composites is not limited to ACL repair. Collagen composites have also been used in the development of tissue analogues including vasculature,4 skin,5 nerve,6 and bone.7

Collagen composites typically undergo gelation via a visco-elastic transition when brought to 37°C8,9 by nucleation of collagen monomers forming branched cross-linked networks.10 Therefore, these solutions are typically stored at 4°C until use to prevent premature network formation. As a result, when these cold solutions are placed into the wound site at 4°C, it can take in excess of 60 min for complete gelation to occur.11 This extensive time requirement makes the use of these materials prohibitive in the clinical arena. Therefore, there has been much recent interest in a controlled heating of the gels to a temperature higher than 4°C before placement in the wound site to lessen the gelation time and increase surgical efficiency.

A few researchers have examined visco-elastic properties (gelation time, elastic modulus [G′], and inelastic modulus [G″]) of collagen type I composites.10,12,13 Forgacs et al.10 investigated the visco-elastic properties of collagen composites correlating phase contrast microscopy and rheologic results, finding that the visco-elastic transition point occurs when collagen nucleation clusters begin to interconnect forming a gel. Newman et al.12 reported that the time to gelation was inversely proportional to the time the neutralized collagen was kept on ice prior to testing. The work by Djabourov et al.13 correlated rheological properties of collagen gels with gelation mechanisms, and suggested that while collagen and gelatin are similar on a molecular level, their respective gelation processes occur though different mechanisms. There has also been an investigation into how the collagen extracted from rats of different ages affects the collagen-composite visco-elastic properties.14 Furthermore, there has been a reasonable amount of research using collagen composites as experimental scaffolds;47,1517 however, few studies have examined their rheologic or visco-elastic properties as a function of temperature. Some research has been completed comparing rheologic data for gels made of collagen to those made of fibrin. The work by Raeber et al.18 shows that cross-linking density varies between the type of gels, and this leads to the differences in mechanical strength. He noted fibrin gels to have greater moduli than collagen gels. Furthermore, work completed by Weisel19 showed that for fibrin gels, increasing fiber diameter resulted in gels with greater mechanical properties. To our knowledge, there are no studies evaluating the mechanical properties of collagen-platelet-plasma composite materials in terms of rheologic behavior, although there have been several studies on the biologic consequences of creating these composites.13,20,21

The hypothesis tested in this article was that increasing the temperature of the gel at the time of injection would significantly decrease the time between injection and gelation (as measured by time to the visco-elastic transition point), but would not significantly alter the mechanical properties of the composite or the subsequent repair tissue. To test this hypothesis, rheologic properties of the composites as a function of temperature were measured in vitro, and then composites in a range of temperatures were placed in a healing wound and the mechanical properties of the resulting structure measured after 14 weeks in vivo. The primary outcome measures were the time to the visco-elastic transition point, the maximum elastic modulus of the composite, and the yield load of an ACL repair after 14 weeks in vivo. Secondary outcome measures included the maximum inelastic modulus and size of the in vivo scar mass as determined by magnetic resonance imaging (MRI).

Materials and Methods

In Vitro Study: Experimental Design

To evaluate the effect of injection temperature on the final rheologic properties of the collagen-platelet gels, composites were manufactured under specified experimental conditions using a specifically designed manufacturing device (Fig. 1) and then injected directly onto the plate of a small oscillation rheometer. Gelation was allowed to progress to completion as defined by a plateau in the elastic modulus values. Changes in elastic and inelastic modulus and time to the visco-elastic transition and the plateau in elastic modulus were recorded for all samples (Fig. 2).

Figure 1
Device for controlled mixing and heating of the collagen composites. The components of the composite were drawn up the syringe using the collapsible auger. The syringe was then placed in the cradle and attached to the motor via a hex connection on the ...
Figure 2
Gelation result for a collagen-platelet composite mixed for 30 s at 100 rpm, and heated at 0.05°/s.

Manufacturing of Acid-Soluble Collagen Used in Composites

The collagen used in this study was derived from rat tails that were obtained from control breeder rats undergoing euthanasia for other Institutional Animal Care and Use Committee approved studies at our institution. The rat-tail tendons were sterilely harvested, minced, and solubilized in 0.01 N hydrochloric acid. The collagen content in the resulting solution was found to be >5 mg/mL. Previous collagen slurries obtained using the same in-house methodology were shown to be made mostly of type I collagen based on their amino acid profile and SDS-PAGE migration pattern. The same collagen solution was used in all experiments.

Prior to testing, the collagen solution was combined with HEPES Buffer (Cellgro; Mediatech, Inc., Herndon, VA), Ham's F-10 medium (MP Biomedicals, LCC, Aurora, OH), antibiotic-antimycotic solution (Cellgro), and sterile water. Sodium bicarbonate (7.5%; Cambrex BioScience Walkersville, Inc., Walkersville, MD) was used to neutralize the acidic mixture to a pH of 7.4. The neutralized collagen solution was kept on ice until use.

Preparation of Platelet Solution

Five hundred milliliters of whole blood was drawn from each of two hematologically normal pigs undergoing other Institutional Animal Care and Use Committee approved studies for the rheologic studies and autologously for each of the five animals in the in vivo studies. Blood was collected in a bag containing 10% by volume acid-citrate dextrose as an anticoagulant and transferred to centrifuge tubes. The blood samples were centrifuged for 6 min at 150 ×g (GH 3.8 rotor, Beckman GS-6 Centrifuge Beckman Coulter, Inc., Fullerton, CA). The supernatant was collected as platelet concentrate. Complete blood counts (CBCs) were measured for the whole blood and platelet concentrate. In vivo, the platelet solution was stored at room temperature for less then 30 min prior to addition to the neutralized collagen.

Preparation of Gels

One milliliter aliquots of the acid soluble collagen were vortexed with buffer to neutralize the solution. This solution aspirated into a syringe containing a collapsible auguer. One milliliter of platelet concentrate was then aspirated into the same syringe. The syringe was affixed in the cradle and the auger engaged with the mixing motor and the gel heated and mixed according to the specified test conditions as detailed below. Mixing speed, mixing time, and heating rate were controlled using a device made for this testing. (Fig. 1; TNCO, Inc., Whitman, MA). An auger was designed to fit inside the 6 cc syringe held in the cradle. This allowed for mixing of the collagen composite components while simultaneously warming the composite. This device had a motor that was coupled to the auger to allow for control of mixing speed and time, and a heating pad under the syringe that allowed for control of heating rate. The device was driven by a custom LabView (Austin, TX) application that allowed for control of the variables, and logging of feedback data.

Composites were created using three different mixing speeds (50, 100, and 150 rpm), three different mixing times (30, 60, and 120 s), and three different heating rates (0.05, 0.10, and 0.15°/s). All combinations of those parameters were tested in triplicate (19 cases total). Control gels were also tested, mixed for 30 s at 100 rpm, without heating. The final temperature of the gels was recorded for all mixing conditions. Additional triplicate gels having an injection temperature of 24°C–26°C, 26°C–28°C, 28°C–30°C, and 30°C–32°C were also tested. The additional gels were prepared by mixing at 100 rpm and heating at 0.10°/s for the time necessary for the gel to reach the specified final temperature.

In Vitro Rheology

Rheological properties of the gels were determined using Cone on Plate Small Amplitude Oscillatory Shear Rheometry using a TA Instruments AR 1000 Rheometer (New Castle, DE). The rheometer was fitted with a 60 mm 1° acrylic cone, and the base plate was maintained at 25°C. For each test, 1 mL of the CPCs was dispensed onto the rheometer plate. The cone was lowered so that the composite was situated in a 38 μm layer between the cone and plate, subjected to a 1% oscillatory. Strain (γo) with an angular frequency of 6.3 rad/s, and the resultant stress (τ(τ)) were recorded. The stress waveform is broken down to one waveform in phase with the oscillatory strain (τ′), and one waveform 90° out of phase (τ″), with the relationship being: τ=τ′+τ′=τo'sinωto″cosωt. The visco-elastic complex modulus (G*) of the gel can be derived from the relationship G*(t) = G′(t) + iG″(t), where G′(t) = τo′/γo and G″(t) = τo″/γo, with the elastic modulus representing the elastic portion of the scaffold, and the inelastic modulus representing the viscous component. Phase angle (δ), which represents the lag between the applied strain and the resultant stress, is determined by the geometric relationship tan(δ) = G″/G′. A value of 45° represents the intersection of G′(t) and G″(t) and is defined as the visco-elastic transition point. Data points for elastic modulus, inelastic modulus, and phase angle were collected at 0.1 Hz, until the rate of change of the increase in elastic modulus was less then 0.1% for three consecutive data points, which was the mathematical definition of the plateau. A sample result is shown in Figure 2.

In Vivo Studies: Experimental Design

Five 30 kg female Yorkshire pigs were used to study the effect of injection temperature on the in vivo performance of the CPCs. Four animals had bilateral ACL transections and for each of these, one side was treated with a suture repair augmented with a collagen sponge containing a CPC injected at a temperature ranging from 28.9°C–32.4°C, while on the contralateral side, the transection was treated with suture repair with the collagen sponge carrier only. A CPC was used because prior in vivo and in vitro studies have shown gradual platelet activation and platelet growth factor release from these CPCs.20,21 In the remaining animal, unilateral surgery was performed with the augmented repair and the contralateral side left as a contemporary intact control. One of the animals developed a postoperative seroma. The seroma was treated with prophylactic antibiotics until complete wound closure was observed on the collagen-platelet side. This knee was excluded from the study. Therefore, there were a total of four knees in the augmented repair group and four knees in the nonaugmented group. All animals were survived to 14 weeks and then underwent MRI evaluation and euthanasia. Knees were immediately harvested and frozen until biomechanical testing. Load to yield, load to failure, maximum linear stiffness, and displacement to failure were measured.

Surgical Procedure

Institutional Animal Care and Use Committee approvals were obtained prior to beginning the study. Five 30 kg female Yorkshire pigs were used anesthetized per an approved protocol. After anesthesia had been obtained, the pigs were weighed and placed in the supine position on the operating room table. Both hind limbs were shaved, prepared with chlorhexidine followed by betadyne paint, and sterilely draped. No tourniquet was used. To expose the ACL, a 4 cm incision was made over the medial border of the patellar tendon. The incision was carried down sharply through the synovium using electrocautery. The fat pad was released from its proximal attachment and partially resected to expose the intermeniscal ligament. The intermeniscal ligament was released to expose the tibial insertion of the ACL. The anterior horns of the medial and lateral menisci attach separately from the intermeniscal ligament and therefore severing this ligament did not lead to gross meniscal instability. A Lachman maneuver was performed prior to releasing the ACL to verify knee stability. Two #1 Vicryl sutures were secured in the distal ACL stump at the middle and distal thirds of the ACL using a modified Kessler stitch. The ACL was transected completely at the junction of the middle and proximal thirds using a #12 blade. No medial collateral ligament (MCL) transection was performed. Complete transection was verified visually and a repeat Lachman maneuver was positive in all knees with no significant endpoint detected after complete transaction (knee was unstable). An absorbable suture anchor (TwinFix AB 5.0 Suture Anchor with DuraBraid Suture [USP#2]; Smith and Nephew, Inc., Andover, MA) was placed at the back of the femoral notch. The knee was irrigated with 500 cc of sterile normal saline to remove all synovial fluid. Once hemostasis had been achieved, a 1 cm ×1 cm collagen sponge was threaded onto the sutures of the suture anchor and slid up into the intercondylar notch. In prior studies, the use of the collagen sponge combined with platelet-rich plasma showed enhanced healing using this technique at 4 weeks.1 The DuraBraid sutures were then tied to the Vicryl sutures previously placed in the ACL stump using maximum manual tension with the knees in resting flexion (approximately 70°, which is 40° short of full extension in these animals). A batch of CPC was mixed by sequentially drawing up equal aliquots of neutralized collagen solution and autologous platelet concentrate into the mixing and heating device and mixing for 1 min at 50 rpm and 0.10°/s, which resulted in injection temperatures between 28.9°C and 32.4°C. The collagen-platelet mixture was then injected to fill the notch and area around the ACL repair (Fig. 3). The knee was left in resting extension and allowed to gel while the identical technique of suture anchor repair was performed with an identical collagen sponge, but without the addition of the CPC. Gelation was observed to have occurred within 10 min of gel injection by manual examination of the joint and by opacification of the composite. In the knees treated with suture repair and collagen aponge only, blood clot was seen to soak the collagen sponge in situ after suture anchor placement. The incisions were closed in multiple layers with absorbable sutures.

Figure 3
Schematic diagram of the suture anchor and repair to the ligament stump, with the collagen sponge in the gap and the collagen-platelet slurry injected around the repair construct.

The animals were not restrained postoperatively, and were allowed ad lib activity. Once the animals recovered from anesthesia, they were permitted to resume normal cage activity and nutrition ad lib. Buprenex 0.01 mg/kg IM once and a Fentanyl patch 1–4 mcg/kg transdermal were provided for postoperative analgesia. All animals were weight bearing on their hind limbs by 24 h after surgery. After 14 weeks in vivo, the animals were again anesthetized and underwent in vivo MR imaging using the protocol detailed below.

After the MR images had been obtained, the animals were euthanized using Fatal Plus at 1 cc/10 lbs. No animals had any surgical complications of difficulty walking normally, redness, warmth, and fever, or other signs of infection that would have necessitated early euthanasia.

Six intact control knees were obtained from age-, gender-, and weight-matched animals after euthanasia following surgical procedures to the chest. The hind limbs were frozen at −20°C for 3 months and thawed overnight at 4°C before mechanical testing. All other testing conditions for these knees were identical to those in the experimental groups.


In vivo MRI was performed at 1.5 Tesla (GE Medical Systems, Milwaukee, WI) with an eight-channel phased array coil at the specified time points. Scanning was performed with the knees placed maximum extension (between 30 and 45° of flexion). Conventional MR included multiplane T1, FSE PD, and T2 weighted images. Field of view (FOV), 16–18 cm matrix; 256×256 (repetition time/echo time) TR/TE, 400/16, 2,500/32, 3,000/66 ms; echo train length (ETL), 8; bandwidth (BW), 15 kHz; slice thickness, 3; interslice gap, 1 mm.

Biomechanical Testing

The bone-ligament-bone ACL complex from both knees for each pig was tested in uniaxial tension as previously described.22 In brief, testing was performed with the knee flexed at 30° of flexion to align the ACL with the load axis. All tissue with the exception of the ACL was resected. Testing was performed at room temperature. Immediately after preconditioning, each specimen was tested to failure in uniaxial tension at 20 mm/min.23,24 Close-range digital images were acquired at 3 Hz using a high resolution digital camera with a macro lens (PixeLINK PLA662 Megapixel Firewire camera; PixeLINK, Ottawa ON, Canada) to determine failure mode. The yield load (defined as the point where the load-displacement curve becomes nonlinear), displacement at yield, tangent modulus (maximum slope of force-displacement curve), maximum load at failure, displacement at failure, and total work to failure (area under force-displacement curve) were determined from the force-displacement curve measured for each bone-ligament-bone ACL complex. The yield load represented the point along the normalized force-displacement curve where the mechanical behavior of the ACL complex departed from “linear” behavior and for the purposes of this analysis was defined as the point where the tangent modulus declined by at least 2% from its maximum value. The displacement at yield was the displacement recorded at this same point. The maximum load is the maximal normalized load sustained by the ACL complex prior to failure and the displacement at failure the displacement recorded at the maximum load. The energy to failure was derived by integrating the total area under the force-displacement curve.

Histological Analysis

After mechanical testing, the knees were fixed in formalin for 1 week, decalcified, and 7 μm thick sagittal sections of the entire knee were stained using hematoxylin and eosin. Digital photographs were taken at 20× of the femoral and tibial insertion sites, and one photograph was taken midsubstanace.

Statistical Analysis

Statistical analysis was performed using SAS StatView 3.0 (SPS Institute, Cary, NC). For the in vitro studies, single factor analysis of variance (ANOVA) was used with the Bonferroni Dunn post-hoc test to compare between groups. A p value <0.05 was deemed statistically significant. For the in vitro studies linear regression were performed, and r2 values and 95% confidence intervals were reported.


In Vitro Results


The centrifugation method used here resulted in a platelet enrichment factor of 2.05 ± 0.42× (mean ± SD). In the platelet solution, the red blood cell count averaged 3.50 × 104 ± 0.70 × 104 cells/μL, and the white blood cell count averaged 2,450 ± 354 cells/μL.


Increasing the mixing time from 30 to 120 s resulted in a 65% increase in the injection temperature (20.29 ± 0.06°C to 33.08 ± 0.95°C, p < 0.001). The time to the visco-elastic transition point decreased 90% from 3.10 ± 0.00 min (30 s cycle) to 0.27 ± 0.03 min (120 s cycle) (p < 0. 001). Increasing the mixing time from 30 to 120 s also resulted in a 95% decrease in the maximum elastic modulus of the gels: 112.25 ± 34.04 Pa to 5.68 ± 1.92 Pa (p < 0.005) (Fig. 4A). Increasing the mixing time had no significant change on the time required to reach the maximum elastic modulus (16.20 ± 2.63 min [30 s cycle], 15.36 ± 2.56 min [60 s cycle], 9.50 ± 5.20 min [120 s cycle], p > 0.01 for all comparisons); however, increasing the mixing time from 60 to 120 s did significantly decrease the time to the maximum inelastic modulus, 15.01 ± 2.62 min to 8.03 ± 6.83 min (p < 0.02).

Figure 4
(A) Mean elastic and inelastic modulus for the collagen-platelet composites as a function of mixing time. † represents a statistically significant difference between 30 and 120 s. * represents a statistically significant difference between 60 ...

Increasing the mixing speed from 50 to 200 rpm did not significantly affect the injection temperature of the gel (23.97 ± 1.00°C to 25.31 ± 0.65°C; p > 0.1). Time to the visco-elastic transition point, maximum elastic and inelastic moduli, and time to the maximum elastic and inelastic moduli were not affected by increasing the mixing speed (p > 0.1).

Increasing the heating rate from 0.05°/s to either 0.10°/s or 0.15°/s significantly increased the injection temperature of the gel: 22.03 ± 0.57°C to 23.97 ± 1.00°C or 25.33 ± 0.97°C respectfully (p < 0.001). Increasing the heating rate from 0.05°/s to either 0.1°/s or 0.15°/s significantly decreased the time to the visco-elastic transition point: 3.63 ± 0.35 min to 2.47 ± 0.42 min or 2.07 ± 0.50 min respectfully (p < 0.002). However, there was no statistically significant change in the maximum elastic (105.90 ± 19.81 Pa, 124.47 ± 35.02 Pa, 88.60 ± 55.82 Pa) or inelastic modulus (24.96 ± 4.18 Pa, 30.56 ± 22.16 Pa, 22.16 ± 15.77 Pa) or the time required to reach these values (p > 0.1) (Fig. 4B).

Gels injected directly from the chilled chamber (average injection temperature 14.41 ± 0.1.24°C) took significantly longer to reach the visco-elastic time point (3.57 ± 0.25 min) compared to the gels heated at 0.10°/s or 0.15°/s (p < 0.05) (average injection temperatures 23.97 ± 1.00°C or 25.33 ± 0.97°C respectively; p < 0.0001 for difference in injection temperature for both comparisons). The maximum elastic and inelastic modulus for the unheated gels injected at 14°C were significantly greater than any of the heated gels (269.50 ± 41.02 Pa and 73.14 ± 13.44 Pa, respectfully) (p < 0.0001), and the time required to reach these values was not significantly changed (12.50 ± 0.27 min for time to maximum elastic modulus and 12.37 ± 0.31 min for time to maximum inelastic modulus) (p > 0.05).

The injection temperature had a significant effect on the gelation time (when defined as the time to the visco-elastic transition point), with gels injected between 20°C and 25°C having the longest gelation time, and a steady decrease in gelation time noted as injection temperature increased (Fig. 5). Similar results were found for gelation time when defined as the time to reach maximum elastic and inelastic modulus.

Figure 5
Mean time to visco-elastic transition point for the collagen-platelet composites as a function of injection temperature. A significant difference was seen between all of the groups. Error bars represent ± one standard deviation.

Both the maximum elastic and maximum inelastic modulus were also affected by injection temperature. There was no significant change in either modulus when comparing the gels injected between 20°C–25°C (111.64 ± 31.08 Pa and 26.88 ± 7.76 Pa) and those injected between 25°C–30°C (120.28 ± 37.51 Pa and 30.20 ± 9.94 Pa;p > 0.1). However, a significant decrease in both moduli was noted when increasing the injection temperature above 30°C (50.18 ± 40.22 Pa and 12.78 ± 9.65 Pa;p < 0.001).

In Vivo Results


The centrifugation method used here resulted in a platelet enrichment factor of 3.9 ± 2.1X (mean ± SD). In the platelet concentrate, the red blood cell count averaged 3.5×104 ± 7.1×103 cells/μL, and the white blood cell count averaged 2,645 ± 115 cells/μL.

Gross Observations and Histology

Upon removal of the capsule and meniscal tissue for mechanical testing of the ACL repair constructs, there was noted to be a fibrovascular band bridging between femur and tibia in the location expected for the ACL in all specimens (Fig. 6B). There was no sign of intact suture material, although in the majority of the specimens, broken remnants of the Durabraid suture were noted. There was no significant synovitis noted in any of the knees, and no gross cartilage or meniscal damage was seen in any of the animals.

Figure 6
(A) Magnetic resonance imaging (MRI) appearance of the healing ACL at 15 weeks after surgery. (B) Repaired ACL ligament 14 weeks following surgery. In all cases, a fibrovascular band bridging between femur and tibia was noted. (C) Histology of repaired ...

Histology of the repair tissue noted in the region of the ACL revealed a fibrovascular scar tissue, characterized by a densely packed collagenous matrix with varying degrees of fiber organization. Near the insertion sites of the ACL, the collagen was in more parallel bundles, and this increased organization likely represented the original tibial and femoral ACL remnants. Bridging between these organized areas was a hypercellular tissue with more disorganized collagen fibers (Fig. 6C) and blood vessels ranging from those with a single cell wall, consistent with capillaries, to those with a more cellular wall, consistent with larger vessels. Red blood cells were observed in the lumen of the vessels, suggesting these vessels were actively perfusing at the time of sacrifice.

Mechanical Properties

Heating the collagen slurry to temperatures approaching body temperature resulted in a mechanically weaker repair tissue. For injection temperatures between 28°C and 33°C, the yield load of the healing ACL was detrimentally affected by increasing the injection temperature of the collagen. Injection temperature was inversely correlated with the strength of the healing ACL after 14 weeks in vivo (r2>0.94; Fig. 7). The strength of these repairs ranged from 88 to 184 N (Table 1, Supplemental Data). In the contralateral knees treated with suture repair with a collagen sponge alone, and without the CPC, the strength of the repair averaged 206 ± 47 N.

Figure 7
The inverse correlation between the injection temperature of the collagen-platelet composite and the resultant yield load of the repaired ACL after 14 weeks in vivo. A similar inverse correlation was observed between injection temperature and the size ...
Table 1
Confidence Interval Data for Observed In Vivo Results

MRI data were also analyzed and showed that the volume and maximum and minimum cross-sectional area of the resulting scar mass were inversely correlated with injection temperature. The scar volume measured 1.1, 0.77, 0.49, and 0.53 cm3 for injection temperatures ranging from 28°C to 33°C (r2 > 0.88) (Table 1, Supplemental Data). Maximum cross-sectional area of the scar mass had areas of 122, 44, 57, and 33 mm2 for the injection temperatures ranging from 28°C to 33°C (r2 > 0.70) (Table 1, Supplemental Data). Minimal cross-sectional area of the scar mass had areas of 51, 34, 20, and 20 mm2 for the injection temperatures ranging from 28°C to 33°C (r2 > 0.91) (Table 1, Supplemental Data). For reference, the cross sectional area of the USP #2 Durabraid sutures was 0.2 mm2 and of the USP #1Vicryl sutures was 0.13 mm2 at the time of surgical repair.


The temperature at the time of injection for CPCs has a significant effect on gel performance in vitro, and preliminary results suggest this is also reflected in the in vivo functional performance of the gel. In the in vitro studies here, injection temperatures of over 30°C resulted in gels with 50% lower stiffness than those with injection temperatures below 30°C. In the in vivo studies, CPCs injected at a higher temperature resulted in a lower yield load of the healing ligament—with a loss of 50% of the repair strength when the temperature increased from 29°C to 32.5°C.

One hypothesis to explain this phenomenon is that as the temperature of the CPC increases, networking of the collagen occurs during the gelation process, and the mechanical perturbation of this network during injection disrupts the nascent collagen quaternary structure. Whether these networks are formed of irreversible covalent crosslinks or the network is simply one of physical aggregation that is subsequently disrupted is unknown.

That at least some of this network may be due to covalent crosslinking is supported by findings in previous studies where increasing the temperature of collagen in solution results in an increased density of crosslink formation,25 as well as studies showing that these crosslinks are irreversible and once disrupted they do not re-form.26 Additional studies have demonstrated that while collagen composites can be quite firm, when stress is applied to them the collagen assembly can easily be broken,11 thus if there is partial gelation in the syringe, the stress of injection through a needle may be enough to disrupt the collagen composite network.

In the in vivo arm of the study, the use of the warmed CPC resulted in an average maximum load of 154 ± 44 N (mean±SD, n=4), a value almost twice that previously reported in a group of animals treated with an identical suture technique with suture alone where the maximum load was 81 ± 43 N (mean ± SD, n=5).27 However, despite this encouraging finding, the group treated with the collagen carrier alone had superior mechanical properties to the group treated with a warmed collagen gel. This suggests that use of gels at temperatures above 28°C may actually be detrimental to wound healing. This may be because that with the elevated temperature of injection used in this study, the gels that form are weaker than the blood clot that would form within the sponge, perhaps so weak that while they initially fill the wound site, they prematurely break apart in the mechanical fluid flow of the synovial environment. A second possibility is that the warmed CPC causes an increased inflammatory response in the knee that leads to a more rapid degradation of the resorbable suture material. However, it may also be that the addition of the CPC at any temperature will also be detrimental. This appears less likely given the encouraging data comparing cold CPC with suture repair alone that reported a doubling of the repair strength with the addition of the cold CPC.1 Additional studies comparing the use of a collagen sponge alone versus a cold CPC are planned and should be useful for dissecting out the mechanism behind this observation.

One of the principal limitations of the in vivo studies is the small number of repairs studied. Only four animals were treated with the heated CPC. The small number of animals was studied only to detect if very large effects of heating the CPC on healing tissue mechanical properties were present. Interestingly enough, even with a sample of only four animals, the correlation coefficient between the mechanical strength of the repairs at 15 weeks and the injection temperature of the composite was very high (r2=0.94). However, given the small number of animals treated with a heated CPC, these results must be considered only a trend, and future studies are required to adequately investigate the effects of heating CPCs on the healing properties of the ACL.

In summary, while increasing the temperature of the CPCs did successfully decrease the required in situ gelation time required for use of these materials clinically, detrimental results in rheologic properties of the gel, and more importantly, in the mechanical properties of the healing tissue, were also observed. These studies suggest that increasing injection temperature of collagen-platelet gels results in a decrease in performance of the gel in vitro and in vivo and this technique should only be used with great caution.


Funding received from CIMIT through DoD funding under cooperative agreement no. DAMD17-02-2-0006 and NIH/NIAMS grant AR049346 and AR052772 (M. M. M.).


Disclosure. Since submission of this article, Martha Murray has become a shareholder and scientific advisory board member for Connective Orthopaedics.


1. Murray MM, Spindler KP, Abreu E, et al. Collagen-platelet rich plasma hydrogel enhances primary repair of the porcine anterior cruciate ligament. J Orthop Res. 2007;25:81–91. [PubMed]
2. Murray MM, Spindler KP, Ballard P, et al. Enhanced histologic repair in a central wound in the anterior cruciate ligament with a collagen-platelet-rich plasma scaffold. J Orthop Res. 2007;25:1007–1017. [PubMed]
3. Murray MM, Spindler KP, Devin C, et al. Use of a collagen-platelet rich plasma scaffold to stimulate healing of a central defect in the canine ACL. J Orthop Res. 2006;24:820–830. [PubMed]
4. Brinkman WT, Nagapudi K, Thomas BS, et al. Photo-cross-linking of type I collagen gels in the presence of smooth muscle cells: mechanical properties, cell viability, and function. Biomacromolecules. 2003;4:890–895. [PubMed]
5. van den Bogaerdt AJ, van Zuijlen PP, van Galen M, et al. The suitability of cells from different tissues for use in tissue-engineered skin substitutes. Arch Dermatol Res. 2002;294:135–142. [PubMed]
6. Joosten EA, Bar PR, Gispen WH. Collagen implants and cortico-spinal axonal growth after mid-thoracic spinal cord lesion in the adult rat. J Neurosci Res. 1995;41:481–490. [PubMed]
7. Wiesmann HP, Nazer N, Klatt C, et al. Bone tissue engineering by primary osteoblast-like cells in a monolayer system and 3-dimensional collagen gel. J Oral Maxillofac Surg. 2003;61:1455–1462. [PubMed]
8. Kadler KE, Hulmes DJ, Hojima Y, et al. Assembly of type I collagen fibrils de novo by the specific enzymic cleavage of pC collagen. The fibrils formed at about 37 degrees C are similar in diameter, roundness, and apparent flexibility to the collagen fibrils seen in connective tissue. Ann N Y Acad Sci. 1990;580:214–224. [PubMed]
9. Sato K, Ebihara T, Adachi E, et al. Possible involvement of aminotelopeptide in self-assembly and thermal stability of collagen I as revealed by its removal with proteases. J Biol Chem. 2000;275:25870–25875. [PubMed]
10. Forgacs G, Newman SA, Hinner B, et al. Assembly of collagen matrices as a phase transition revealed by structural and rheologic studies. Biophys J. 2003;84:1272–1280. [PubMed]
11. Wallace DG, Rosenblatt J. Collagen gel systems for sustained delivery and tissue engineering. Adv Drug Deliv Rev. 2003;55:1631–1649. [PubMed]
12. Newman S, Cloitre M, Allain C, et al. Viscosity and elasticity during collagen assembly in vitro: relevance to matrix-driven translocation. Biopolymers. 1997;41:337–347. [PubMed]
13. Djabourov M, Lechaire JP, Gaill F. Structure and rheology of gelatin and collagen gels. Biorheology. 1993;30:191–205. [PubMed]
14. Wu CC, Ding SJ, Wang YH, et al. Mechanical properties of collagen gels derived from rats of different ages. J Biomater Sci Polym Ed. 2005;16:1261–1275. [PubMed]
15. Mio T, Adachi Y, Romberger DJ, et al. Regulation of fibroblast proliferation in three-dimensional collagen gel matrix. In Vitro Cell Dev Biol Anim. 1996;32:427–433. [PubMed]
16. Lin PW, Wu CC, Chen CH, et al. Characterization of cortical neuron outgrowth in two- and three-dimensional culture systems. J Biomed Mater Res B Appl Biomater. 2005;75:146–157. [PubMed]
17. Yang B, Cao DJ, Sainz I, et al. Different roles of ERK and p38 MAP kinases during tube formation from endothelial cells cultured in 3-dimensional collagen matrices. J Cell Physiol. 2004;200:360–369. [PubMed]
18. Raeber GP, Lutolf MP, Hubbell JA. Molecularly engineered PEG hydrogels: a novel model system for proteolytically mediated cell migration. Biophys J. 2005;89:1374–1388. [PubMed]
19. Weisel JW. The mechanical properties of fibrin for basic scientists and clinicians. Biophys Chem. 2004;112:267–276. [PubMed]
20. Jacobson M, Fufa D, Abreu EL, et al. Platelets, but not erythrocytes, significantly affect cytokine release and scaffold contraction in a provisional scaffold model. Wound Repair Regen. 2008;16:370–378. [PMC free article] [PubMed]
21. Fufa D, Shealy B, Jacobson M, et al. Activation of platelet-rich plasma using soluble type I collagen. J Oral Maxillofac Surg. 2008;66:684–690. [PMC free article] [PubMed]
22. Spindler KP, Murray MM, Devin C, et al. The central ACL defect as a model for failure of intra-articular healing. J Orthop Res. 2006;24:401–406. [PubMed]
23. Katsuragi R, Yasuda K, Tsujino J, et al. The effect of nonphysiologically high initial tension on the mechanical properties of in situ frozen anterior cruciate ligament in a canine model. Am J Sports Med. 2000;28:47–56. [PubMed]
24. Sakai T, Yasuda K, Tohyama H, et al. Effects of combined administration of transforming growth factor-beta1 and epidermal growth factor on properties of the in situ frozen anterior cruciate ligament in rabbits. J Orthop Res. 2002;20:1345–1351. [PubMed]
25. Nickerson MT, Patel J, Heyd DV, et al. Kinetic and mechanistic considerations in the gelation of genipin-crosslinked gelatin. Int J Biol Macromol. 2006;39:298–302. [PubMed]
26. Stenzel KH, Miyata T, Rubin AL. Collagen as a biomaterial. Annu Rev Biophys Bioeng. 1974;3:231–253. [PubMed]
27. Murray MM, Palmer M, Abreu E, et al. Platelet-rich plasma alone is not sufficient to enhance suture repair of the ACL in skeletally immature animals: an in vivo study. J Orthop Res. in press. [PMC free article] [PubMed]