MRI of the proximal femur predicts marrow cellularity and the number of mesenchymal stem cells

Authors

  • Kuen Tak Suh MD, PhD,

    Corresponding author
    1. Department of Orthopaedic Surgery, Pusan National University School of Medicine, Yangsan-city, Gyeongsangnam-do, Korea
    2. Medical Research Institute of Pusan National University, Pusan National University School of Medicine, Yangsan-city, Gyeongsangnam-do, Korea
    • Department of Orthopedic Surgery, Pusan National University School of Medicine, Beomeo-ri, Mulgeum-eup, Yangsan-city, Gyeongsangnam-do 626-700, Korea
    Search for more papers by this author
  • Jae Min Ahn MD,

    1. Department of Orthopaedic Surgery, Pusan National University School of Medicine, Yangsan-city, Gyeongsangnam-do, Korea
    Search for more papers by this author
  • Jung Sub Lee MD, PhD,

    1. Department of Orthopaedic Surgery, Pusan National University School of Medicine, Yangsan-city, Gyeongsangnam-do, Korea
    2. Medical Research Institute of Pusan National University, Pusan National University School of Medicine, Yangsan-city, Gyeongsangnam-do, Korea
    Search for more papers by this author
  • Jung Yun Bae MD,

    1. Department of Orthopaedic Surgery, Pusan National University School of Medicine, Yangsan-city, Gyeongsangnam-do, Korea
    Search for more papers by this author
  • In Suk Lee MD,

    1. Department of Radiology, Pusan National University School of Medicine, Pusan, Korea
    Search for more papers by this author
  • Hak Jin Kim MD, PhD,

    1. Medical Research Institute of Pusan National University, Pusan National University School of Medicine, Yangsan-city, Gyeongsangnam-do, Korea
    2. Department of Radiology, Pusan National University School of Medicine, Pusan, Korea
    Search for more papers by this author
  • Jin Sup Jung MD, PhD

    1. Medical Research Institute of Pusan National University, Pusan National University School of Medicine, Yangsan-city, Gyeongsangnam-do, Korea
    2. Department of Physiology, Pusan National University School of Medicine, Yangsan-city, Gyeongsangnam-do, Korea
    Search for more papers by this author

Abstract

Purpose:

To determine whether the marrow conversion index (MCI) in MRI is related to the total number of mononuclear and mesenchymal stem cells (MSCs) in proximal femoral metaphysis of patients with hip osteoarthritis.

Materials and Methods:

Thirty-two hips of 32 consecutive patients who underwent total hip arthroplasty (THA) for hip osteoarthritis were included in this study. MRI of the hip was performed preoperatively and MCI was subsequently calculated. Three-milliliter bone marrow samples were obtained from the proximal femur during THA and the number of total mononuclear cells was determined using a hemocytometer. Colony forming unit-fibroblasts (CFU-Fs) assays of MSCs were performed by transferring a total of 2 × 104 mononuclear cells to each of five 60-mm plates. One week later, the numbers of colonies were counted.

Results:

The total number of mononuclear cells decreased with increasing MCI. Likewise, the prevalence and total number of CFU-Fs increased with increasing number of total mononuclear cells, and decreased with increasing MCI.

Conclusion:

Our results suggest that measurement of MCI in MRI can be an objective and noninvasive method to predict marrow cellularity and the number of MSCs in patients with hip osteoarthritis. J. Magn. Reson. Imaging 2012;35:218-222. © 2011 Wiley Periodicals, Inc.

HEMATOPOIETIC MARROW DECREASES progressively with age, corresponding to an increase in fatty marrow (1). Until now, marrow cellularity, one of the few quantifiable changes during this transition, is usually estimated by a bone marrow biopsy from the iliac crest or sternum; however, there are many limitations associated with this approach (2). For example, latex gloves, local anesthetic, and use of antiseptics or sedatives can cause allergic reactions. Furthermore, it may not be feasible for a single clinician to perform the aspiration of bone marrow. Excessive bruising or bleeding is also likely in patients with thrombocytopenia or abnormal coagulation tests. Thus, as an alternative to direct methods such as bone marrow biopsy, several MRI studies assessing the ability to measure the abundance of marrow cells have been reported (3–6). In these studies, MRI was used to estimate marrow composition because it provides quantitative information concerning the amount of fat in bones. Proximal femoral metaphysis is one of the predominant sites in which hematopoietic marrow is macroscopically identifiable during adult life. On the other hand, the marrow of the greater trochanter becomes obviously fatty before puberty (1). Due to the presence of these features, it is possible to compare the marrow of the proximal femoral metaphysis with the marrow of the greater trochanter using a single imaging sequence. Indeed, the fatty marrow conversion of the proximal femoral metaphysis can be objectively and reliably evaluated by measuring the marrow conversion index (MCI). The MCI can be calculated by dividing the signal intensity of the proximal femoral metaphysis by that of the greater trochanter (7). The MCI depends not only on the distribution of the hematopoietic and fatty marrow, but also on the signal intensity of the hematopoietic marrow, that is, hematopoietic marrow cellularity (8). Therefore, the MCI functions as an MRI-derived value of marrow cellularity.

Mesenchymal stem cells (MSCs) regulate hematopoiesis by means of cytokine production in the hematopoietic niche (9). MSCs appear to regulate the survival, self-renewal, migration, and differentiation of hematopoietic stem cells (10). Osteoblasts play a crucial role in hematopoiesis and are a key element of the hematopoietic niche (11). The number of hematopoietic stem cells increases when the number of osteoblasts increases (11). Based on these observations, we hypothesized that the MCI may closely represent actual marrow cellularity, and, furthermore, may be used to predict the total number of MSCs.

To the best of our knowledge, there have been no previous reports that have evaluated the correlation between MRI values and the total numbers of mononuclear cells and MSCs in the proximal femoral metaphysis. Thus, in the present study, we evaluated whether the MCI in T1-weighted MRI might be related to the total numbers of mononuclear cells and MSCs in the proximal femoral metaphysis of patients with hip osteoarthritis.

MATERIALS AND METHODS

Patients

This prospective study was approved by the Human Research Committee of the authors' institution. A total of 32 consecutive patients who visited our institution between April 2005 and July 2007 with hip osteoarthritis were included in this prospective study. Osteoarthritis was confirmed by radiography and MRI of the hip was performed preoperatively. After written informed consent was obtained from all patients, bone marrow samples were procured from the proximal end of the femur while inserting the tapered awl into the femoral canal during total hip replacement surgery. Twelve hips were male and 20 hips were female. Mean age of all the patients was 54.3 ± 10.8 (mean ± standard deviation) years with a range of 34–72 years. The mean time between MRI and biopsy was 17.3 ± 21.78 days (range, 1–70 days).

Measurement of MCI in MRI

MRI of the hip was performed preoperatively with a 1.5T superconducting magnet (Magnetom or Sonata; Siemens, Erlangen, Germany). Coronal T1-weighted spin-echo images were obtained in all hips with the following parameters using a body coil: repetition time (TR) of 420–596 ms, echo time (TE) of 20 ms, 256 × 256 matrix, two averages, 5-mm section thickness, 0- to 1-mm intervening gap, and 34- to 40-cm field of view. Patients with diffuse marrow edema in the proximal femur were excluded from our study. All measurements were performed by one musculoskeletal radiologist (I.S.L.) who was blinded to the clinical and laboratory data. The marrow signal intensities of the proximal femoral metaphysis and the greater trochanter on T1-weighted images were measured with the Sonata Evaluation Package (Siemens) by using the cursor to define the regions of interest from the midsection of the proximal femur in the coronal plane (Fig. 1). The region of interest of the proximal femoral metaphysis was defined as follows: the superior margin was the intertrochanteric line of the femur, the superolateral margin was the physeal scar between the proximal femoral metaphysis and the greater trochanter, the distal margin was at the same level of the lower end of the lesser trochanter, and the mediolateral margin was the inner surface of the mediolateral cortices excluding the lesser trochanter. MCI was calculated as follows: MCI (%) = (signal intensity of the proximal femoral metaphysis ÷ signal intensity of the greater trochanter) × 100 (7, 8).

Figure 1.

Region of interest of the proximal femur on T1-weighted image. T1-weighted coronal images were obtained with the following parameters: TR = 420–596 ms, TE = 20 ms, 256 × 256 matrix, two averages, 5-mm section thickness, 0- to 1-mm intervening gap, and 34- to 40-cm field of view. The region of interest of the proximal femur was defined by using the cursor from the midsection of the proximal femur in the coronal plane.

Separation of Total Marrow Mononuclear Cells

Three-milliliter fresh bone marrow samples were obtained from the proximal end of the femur while inserting the tapered awl into the femoral canal during total hip arthroplasty surgery. Mononuclear cells from the bone marrow were separated by centrifugation in a Ficoll-Hypaque gradient (density = 1.077 g/cm3; Sigma, St. Louis, MO) and then suspended in α-modified Eagle's medium containing 10% fetal bovine serum, 100 U/mL of penicillin and 100 μg/mL of streptomycin. The total number of mononuclear cells was then counted with a hemocytometer and values were expressed as the number of mononuclear cells (in millions) per mL.

Colony Forming Unit-Fibroblasts (CFU-Fs) Assays

For CFU-F assays, a total of 2 × 104 mononuclear cells were transferred to each of five 60-mm plates, which were incubated at 37°C in a humidified atmosphere containing 5% CO2. One week later, the number of colonies was counted manually to quantify the prevalence of CFU-Fs in each of the five plates. The sum of CFU-Fs isolated per 2 × 104 mononuclear cells plated from five individual plates was used to determine the prevalence of CFU-Fs (colonies/105 cells). To determine the reproducibility of the CFU-F experiments, we determined CFU-Fs in triplicate using two different samples. Triplicate determinations of CFU-F numbers for each sample were reproducible according to our experimental conditions.

Calculation of the Total Number of CFU-Fs

The total number of CFU-Fs per mL can be calculated as the product of the total number of mononuclear cells obtained per mL of aspirate and the sum of CFU-Fs per 2 × 104 mononuclear cells from the five plates assayed. Thus, in our study, the total number of CFU-Fs per mL was calculated according to the following equation: total number of CFU-Fs per mL (colonies/mL) = (the total number of mononuclear cells) × (the sum of CFU-Fs count in five plates) = (a × 106 mononuclear cells/mL of the bone marrow sample) × (b × colonies of CFU-Fs/105 mononuclear cells) = 10 × a × b CFU-Fs/mL.

Statistical Analysis

Values are expressed as the mean ± standard deviation. Nonparametric tests were used for statistical analysis because of small sample sizes. Results were analyzed with Spearman's correlation test using SPSS version 14.0 software (SPSS Inc., Chicago, IL).

RESULTS

Details of the data obtained from the 32 subjects are shown in Table 1, which includes individual data, mean of the MCI, total number of mononuclear cells, prevalence, and total number of CFU-Fs. The mean MCI was 88.8 ± 5.1% , while the number of total mononuclear cells was 1.76 ± 1.85 × 106 cells/mL. The prevalence of CFU-Fs was 71.4 ± 28.9 colonies/105 cells, and the total number of CFU-Fs was 1444 ± 1781 colonies/mL.

Table 1. Details of Data from 32 Subjects
CaseAge/ GenderInterval between MRI and biopsy (days)Marrow conversion index (%)Total number of mononuclear cells (×106/ml)Prevalence of CFU-Fs (colonies /105 cells)Total number of CFU-Fs (colonies/ml)
161 / F286.02.50862150
251 / F187.22.50651625
343 / M281.06.25996188
443 / M1681.75.25874568
558 / M194.72.00611220
669 / F2291.20.5425135
735 / M177.16.25704375
858 / M187.83.70843108
969 / M183.06.401016464
1072 / M398.30.8525213
1160 / M592.51.3054702
1234 / F198.80.461046
1361 / F898.71.0826281
1464 / M1287.11.1485969
1568 / F288.62.381533641
1648 / F794.10.8074592
1746 / M4195.20.6775503
1846 / F2285.20.5066330
1936 / M3383.93.14842638
2045 / F285.60.5755314
2149 / M5686.51.241021265
2245 / F187.40.6055330
2355 / F3784.20.8078624
2471 / F591.90.6899673
2554 / F5388.20.6067402
2649 / M7091.80.50101505
2745 / F1588.00.5080400
2867 / M692.10.5291473
2954 / F6287.70.5523127
3061 / F388.60.6064384
3161 / F6088.50.7876593
3260 / F288.30.6063378

The prevalence and the total number of CFU-Fs increased significantly with increasing number of total mononuclear cells (r = 0.383 and P = 0.030; r = 0.853 and P < 0.001, respectively). Likewise, the total number of mononuclear cells decreased with increasing MCI. The total number of mononuclear cells showed a statistically significant correlation with the MCI (r = −0.440, P = 0.012) (Fig. 2). The prevalence and the total number of CFU-Fs decreased with increasing MCI. The prevalence and the total number of CFU-Fs showed a statistically significant correlation with the MCI (r = −0.386 and P = 0.029, r = −0.529 and P = 0.002, respectively) (Figs. 3 and 4).

Figure 2.

Correlation of the total number of mononuclear cells and total number of CFU-Fs with the marrow conversion index (MCI). The total number of mononuclear cells decreased with increasing MCI (r = −0.440; P = 0.012).

Figure 3.

The prevalence of CFU-Fs decreased with increasing MCI (r = −0.386; P = 0.029).

Figure 4.

The total number of CFU-Fs decreased with increasing MCI (r = −0.529; P = 0.002).

DISCUSSION

We attempted to predict the actual total numbers of mononuclear cells and MSCs using the MCI as an MRI-derived value of marrow cellularity of the proximal femoral metaphysis unlike the invasive method such as marrow biopsy or aspiration. Accurate measurement of marrow cellularity by this approach is now possible due to advances in current imaging methods. Indeed, to the best of our knowledge, there have been no previous reports describing a correlation between MRI signal intensity and the total number of mononuclear cells and MSCs of the proximal femoral metaphysis. In our study, the prevalence and total number of CFU-Fs were significantly increased with an increasing number of total mononuclear cells, and this finding is consistent with the results reported by Majors et al (12). The number of total mononuclear cells, the prevalence of CFU-Fs, and the total number of CFU-Fs were likewise significantly decreased with increasing MCI. Therefore, the MCI may represent an approximate measure of marrow cellularity and predict the number of MSCs.

MSCs form characteristic adherent, fibroblast-like colonies when grown in vitro in fetal calf serum. The most well studied and accessible source of MSCs is bone marrow, although even in this tissue such cells are present at a low frequency. Long-term hematopoietic stem cells have been found attached to MSCs lining the bone surface, supporting the conception that regulation of the size of the osteoblast niche controls the magnitude of hematopoiesis (11). MSCs are precursors of the bone marrow microenvironment and regulate hematopoiesis by means of cytokine production in the hematopoietic niche (9). Therefore, increased hematopoietic marrow is associated with an increased number of MSCs present in the marrow. During in vitro culture, MSCs form distinct colonies of cells with fibroblast morphology, giving rise to the term colony forming unit-fibroblasts (CFU-Fs). Each CFU-F represents the descendants of a single MSC and thus the number of CFU-Fs is a direct estimate of the number of MSCs present at the start of the culture (13). Therefore, the CFU-F assay is an original test for mesenchymal cells, allowing for evaluation of the number of MSCs present in the bone marrow as well as their proliferative abilities (14, 15). Although not all cells that form CFU-Fs represent true progenitor or stem cells, the frequency of CFU-Fs does correlate with the incidence of progenitors in a given bone marrow sample (16).

In conclusion, our results suggest that measuring the MCI in MRI can be an objective and noninvasive method to predict bone marrow cellularity and the number of MSCs of the proximal femoral metaphysis of patients with hip osteoarthritis. Our results provide baseline information on the use of marrow cellularity and MSCs in various fields and also provide a baseline for predicting alteration of marrow cellularity and the number of MSCs using the MCI. Further studies are necessary to determine the clinical relevance of MCI measurements.

Ancillary