Accordingly, we used the PC3N-A6-WT and PC3N-A6-RR prostate cancer cells that express equivalent levels of the wild-type A6 subunit (cleavable to A6p via uPA treatment) and the non-cleavable subunit, respectively (). The surface expression levels of the receptors were equivalent as determined by Flow Cytometry (FACS) analysis (). The ability to cleave the receptor by Urokinase-type Plasminogen Activator (uPA) and the generation of the A6p structural variant is shown schematically (). We next tested the functional properties of A6 integrin using a laminin 111 containing matrix, matrigel, modified to contain laminin 332. Matrigel is a laminin rich extracellular matrix that models physiologically relevant conditions
. The PC3N-A6-WT cells migrated within matrigel in a manner that was integrin dependent (). The cells containing the uncleavable receptor, PC3N-A6- RR, were unable to migrate in matrigel, consistent with previous results using routine tissue culture conditions ()
Biochemical and migration phenotype of PC3N-A6-WT and PC3N-A6- RR cells expressing the wildtype(cleavable) and RR(uncleavable) integrin A6, respectively.
In order to determine the effect of human prostate cancer cells within the bone, we adapted the Clohisy-Mantyh murine model in which cancer cells are directly injected and sealed into the femur of a mouse
. Male SCID mice were anesthetized with ketamine/xylazine and an arthrotomy was performed exposing the condyles of the distal femur as previously described
. A hole was drilled into the femur for the injection needle to ensure accurate placement of tumor cells within the bone. The exact placement of the needle into the intramedullary space of the femur was confirmed by imaging (). Human prostate tumor cells were injected into the right leg of the mouse and the injection site sealed with dental amalgam ().
SCID Mouse xenograft model and quantification of bone destruction after injection of prostate cancer cells into the femoral intramedullary space.
We next determined the deleterious effects of tumor cells residing within the bone using standard radiographic imaging in live animals 21 days following surgery. All animals injected with PC3N-A6-WT, and PC3N-A6-RR cells developed bone loss 21 days following surgery (). Images consistently showed osteolytic activity, particularly of the metaphyseal bone at the distal (knee) end of the femur. Mice injected with the PC3N-A6-WT cells showing dramatically more bone loss compared to those injected with the PC3N-A6-RR cells. No bone loss was observed in animals injected with media alone (). Animals injected with the PC3N-A6-WT cells showed increased bone loss compared to those injected with PC3N-A6-RR cells (). The radiographs were rated according to a 4 point scale in which 0 indicates normal bone and 3 indicates full thickness bicortical bone loss (). Animals injected with PC3N-A6-WT cells showed a dramatic increase in fractures (unicortical or bicortical) 21 days following surgery compared to the PC3N-A6-RR treated mice (). These data indicate that the placement of tumor cells within the bone containing a non-cleavable A6 integrin results in a significant delay in the development of bone loss.
Although the imaging results provided information about severe bone destruction, it did not give direct information about the distribution of the tumor cells. Bone loss is a consequence in part of tumor cells resident within the bone; an event that dramatically affects the critical balance of osteolytic and osteoblastic activity
. We directly investigated the distribution of the tumor cells within the bone using histological analysis by hemotoxylin and eosin staining of decalcified specimens. The mouse bones were carefully oriented for longitudinal sectioning to include observing the epiphyseal plate as well as the distal region of the bone in the same section (). The analysis of the bones from the tumor injected animals demonstrated presence of tumor in the entire intramedullary space of the bone in those injected with PC3N-A6-WT cells along with invasion into the cortical bone (). The tumor cells containing the cleavable integrin had reached the epiphyseal plate; the bone marrow was completely replaced. This result is consistent with the knowledge that once prostate cancer has established itself in bone marrow it will eventually replace the marrow, interrupting bone homeostasis
. In contrast, PC3N-A6-RR injected animals contained tumor cells in the mid-shaft region of the bone and the tumor failed to reach the epiphyseal plate. Normal bone marrow was present in the areas that did not contain tumor cells (). The tumor cells within the mid shaft region or those that had reached the epiphyseal plate were viable and morphologically indistinguishable. (). The tumor distribution pattern was found to be consistent in the histological analysis of all the test animals assayed.
Histological examination of bone destruction, tumor cell distribution and verification of mutated integrin expression after injection of prostate tumor cells.
In order to confirm that the injected tumor cells were expressing the mutated integrin, 21 days following injection of PC3N-A6-WT or PC3N-A6-RR cells, the marrow was expressed from the intramedullary space of the mouse femurs. Analysis of bone marrow samples resulted in mRNA specific for the PC3N-A6-RR cells in marrow from mice injected with PC3N-A6-RR, but not PC3N-A6-WT mice. These data indicated that PC3N cells transfected with the uncleavable A6 integrin maintained expression of the mutant integrin within the intramedullary space of the femur and were present 21 days following injection of the cells into the femur ().
Behavioral analyses of spontaneous and evoked pain were determined 21 days following injection of PC3N-A6-WT or PC3N-A6-RR cells into the femur to evaluate the role of cleavage of A6 integrin on the development of spontaneous and evoked cancer pain behaviors. Spontaneous pain was measured by assessing flinching of the cancer treated hind limb as previously described
. Mice injected with PC3N-A6-WT cells showed increased spontaneous flinching behavior compared to PC3N-A6-RR treated mice which demonstrated low levels of flinching that were comparable to control animals (). Evoked pain, as indicated by tactile allodynia, was determined by paw withdrawal from probing of the hind paw ipsilateral to the cancer treated femur with calibrated von Frey filaments as previously described
. Mice injected with the PC3NA6-WT cells showed tactile allodynia as indicated by lowered threshold for paw withdrawal from von Frey filaments (). In contrast, PC3N-A6-RR injected mice did not show tactile allodynia, with paw-withdrawal thresholds similar to control animals (). Movement-evoked pain was determined by rating the limb use during normal ambulatory movement as previously described
. Mice injected with PC3N-A6-WT cells developed movement evoked pain whereas mice injected with PC3N-A6-RR cells showed no movement-evoked pain, with limb use ratings the same as control animals ().
Development of cancer-induced spontaneous pain, tactile allodynia, and movement-evoked pain in animals 21 days after surgery.
These data indicate that cleavage of integrin A6 influenced the development of cancerinduced spontaneous and evoked pain associated with bone loss and fracture. Preventing cleavage of integrin A6 dramatically reduced bone loss and development of cancerinduced pain. In contrast, expression of the cleavable A6 integrin increased cancerinduced bone loss and fracture with a concurrent increase in pain behavior.