In this study, we used HPE provided by Melsmon (Melsmon, Tokyo, Japan). HPE is a liquid agent prepared by freezing the villous tissue part of fresh placenta and then extracting and purifying it by means of hydrochloric acid hydrolysis. The extract contains 100 mg of water-soluble substances in 2 ml of normal saline solution. The identified components are nucleic acid related substances such as uracil, adenine, guanine and thymine, amino acids such as lysine, alanine, asparagic acid, leucine, glutamic acid, glycine, valine, serine, tyrosine, phenylalanine, threonine, arginine, proline, cystine, methionine, histamine, and minerals such as sodium, potassium, magnesium, phosphorus, iron. In this study, a stock solution of the HPE agent was diluted with normal saline and injected intraperitoneally.

For the preparation of the KBx/N serum transfer arthritis model, KRN TCR transgenic mice provided by D. Mathis and C. Benoist (Harvard Medical School, Boston, USA) were first mated with non-obese diabetic (NOD) mice. Arthritis began to occur about four weeks after birth in KBx/N mice. Blood samples were taken in the eighth week when the arthritis was most severe. Serum was separated by centrifugation and stored frozen. BALB/c male mice (Orient Bio Inc., Sungnam, Korea) aged 7-9 weeks were purchased and grown. KBx/N serum (150 ul) was intraperitoneally injected into BALB/c male mice two times on Day 0 and Day 2 to induce arthritis. Experimental animals were divided into three groups: Group 1 (n=10) was the control group, to which vehicle was administered; Group 2 (n=10) received 1 ul of HPE; Group 3 (n=10) received 100 ul of HPE. Administration of 1 ul HPE was comparable to a dose of 2 ml, which is the recommended therapeutic dose for humans. HPE 100 ul was the maximum dose that did not lead to any deaths or fatal toxicity in the mice in a preliminary test, and we gave it as a one-time dose in this experiment. Intraperitoneal injection of 1 ul or 100 ul HPE and vehicle was given two times per week to all mice, from Day 0 to day 14.

For the preparation of the CIA model, DBA/1 male mice (Orient Bio Inc., Sungnam, Korea) aged 7-9 weeks were purchased. Bovine-type type II collagen (Chondrex inc., Redmond, USA) (150 µg) and the same quantity of complete Freund's adjuvant (CFA) (Chondrex inc., Redmond, USA) were injected intradermically through the tail. Twenty one days later, bovine-type type II collagen (150 µg) and the same quantity of incomplete Freund's adjuvant (IFA, Chondrex) were injected through the tail to induce arthritis. The experimental animals were divided into two groups. Group 1 (n=8), which was the control group, was administered vehicle; Group 2 (n=8) received 1 ul of HPE. Intraperitoneal injection of 1 ul HPE and vehicle was given two times per week to all mice from Day 21 to the end of the experiment.

The incidence and severity of arthritis were assessed macroscopically by observing mice two or three times per week. The severity of arthritis was coded as Grades 0 to 3 for each leg: Grade 0, without any edema or swelling; Grade 1, mild edema and flare localized to the toes or ankle joints; Grade 2, moderate edema and flare over the ankle joints and toes; and Grade 3, severe edema and flare over the region between the ankle joints and the entire toes or accompanied by joint stiffness. The assessment was performed with reference to the sum of the severity measured from the four legs, and thus the maximum possible arthritis score was 12 points. The presence of arthritis was defined as definite arthritis accompanying flare and edema together. The percentage (%) of the legs where arthritis occurred among all the legs of the subject mouse was recorded as the incidence of arthritis. The thickness of the two rear feet was measured at the thickest part of the ankles for all mice using electronic calipers (Orient Bio Inc., Sungnam, Korea). To minimize rater error, arthritis assessment was done by two raters who were not aware which were the experimental and which the control groups.

A mixture of Zoletil (Virbac, Magny-en-Vexin, France) and Rumpun (Bayer, Seoul, Korea) was diluted ten times with phosphate buffered saline (PBS) and 200 ul of the solution was intraperitoneally injected into the mice to anesthetize them. Blood samples were then taken by cardiocentesis. Serum was separated by centrifugation of the blood samples and stored at -70℃ before the test. Ankle joint tissues were taken after removing the skin, rapidly cooling them with liquid nitrogen, and then storing them frozen at -70℃ until use. Ankle tissues were ground in a mortar using liquid nitrogen and proteins were extracted with a lysis buffer (Mammalian Protein Extraction Reagent) (Pierce Biotechnology Inc., Rockford, USA) containing proteinase inhibitor cocktail (Calbiochem, San Diego, USA). In the serum and joint tissue extractions, the cytokine concentrations of TNF-α, IL-1β, IL-6, soluble receptor activator of NF-κB ligand (sRANKL) and IL-10 were measured using ELISA kits (PeproTech inc., Rocky Hill, USA). The capture antibody for the cytokine to be measured was put into a 96-well plate (Nunc, Rochester, USA) for sandwich ELISAs and reacted overnight at 4℃. Then, PBS containing the blocking solution (1% bovine serum albumin (BSA)/0.05% Tween-20) was added and reacted for two hours at room temperature. The standard samples and the samples to be measured were reacted for two hours at room temperature. After washing them with water, the detection antibody was added and reacted for two hours at room temperature. Then, Avidin-peroxidase (50 µl, 1 : 2,000) and ABTS substrate solution were added to each of the wells one by one and allowed to react. Following the reaction, light absorbance was measured at 450 nm with an ELISA reader. a standard curve was acquired using the concentrations and absorbances of standard samples. Then the absorbance of the tested samples converted to concentrations using the standard curve. The measurement was repeated for all the tested samples according to the directions of the manufacturer, and the mean values were used for the analysis.

We performed radiologic assessment to investigate the effect of HPE on the degree of bone destruction. Radiological assessments were performed only with the CIA model because the KBx/N model, which is a two-week model, did not show a clear change that could be found by general radiological tests since destruction of bone was relatively weak.

The radiography was performed at the knee and ankle joints of mice 42 days after the induction of arthritis, using a digital mammograph (Mammomat Novation^DR) (Siemens, Erlangen, Germany). A full field flat panel digital detector that could create a 24×29 cm image (maximum matrix size, 3,328×4,096; pixel size, 70 µm) was used, and all the images were taken under exposure conditions of 30 kVp, 90 mAs and 1.5 magnification. The destruction of the joints was assessed with respect to the region between the patella and the femur and the region between the femur and the tibia in case of knee joints; and with respect to the three regions of the ankle, the intertarsal part and the metatarsophalangeal joint in the case of rear foot joints. Joints received scores of Grade 0 (normal) to grade 5 (complete disruption of the joint) (Table 1). Mean radiological scores were determined as the mean of the measurements at the respective regions.

Histopathologic assessment was performed 14 days after the induction of arthritis in the KBx/N model and 42 days after induction in the CIA model.

For the histopathologic assessment, we took knee and ankle joints and prepared paraffin blocks through a fixation procedure with formalin and a decalcification procedure with Calci-Clear Rapid (Diagnostics, Atlanta, USA). Blocks were sectioned into slices of 4 µm thickness and attached to slides. After deparaffinizing with xylene three times, for five minutes each time, we washed the slices with distilled water and performed hematoxylin and eosin (H&E), Safranine-O and tartate-resistant acid phosphatase (TRAP) staining. The progress of arthritis was assessed quantitatively with respect to bone erosion, pannus formation and cartilage damage. Scores included in five grades between 0 (normal) and 4 (severe) (Table 2).

In 10 randomly selected zones at 200 X magnification, the number of giant cells containing at least three TRAP-positive nuclei was counted, summed and marked as the total number of TRAP-positive cells.

For comparison of mean values among the groups, a Kruskal-Wallis test was performed in the KBx/N model and a Mann-Whitney U-test in the CIA model. All the results were expressed as 'mean±standard deviation' Statistical analysis of the data was done using SPSS Version 11.0 (SPSS, Chicago, USA). Differences where p was less than 0.05 were considered significant.

There was no significant difference in the incidence of arthritis among the three treatment groups (vehicle, HPE 1 ul, HPE 100 ul). The respective arthritis scores were 9.11±1.96, 10.22±1.71, and 10.22±1.48 for the first week and 5.77±1.30, 6.89±1.90, and 5.89±0.78 in the second week (Fig. 1-A, B). There was also no between group differences in arthritis severity. The thickness of the two ankles measured with the electronic calipers in each of the mice was 2.61 mm, 2.61 mm, and 2.60 mm in the initial stage, 3.59 mm, 3.58 mm, and 3.65 mm in the first week, and 3.06 mm, 3.17 mm, and 3.19 mm in the second week in the respective groups (Fig. 1-C). No significant difference was found in the improvement or exacerbation of arthritis due to HPE. Neither decreased weight nor death was found in the 100 ul HPE group (Fig. 1-D).

For histological assessment, we performed H&E and Safranine-O staining, and quantitatively assessed disease progress in terms of synovial inflammation, cartilage damage, and bone erosion. In the vehicle, HPE 1 ul, and HPE 100 ul groups, the infiltration of synovial inflammation cells was 2.83±0.24, 2.36±0.29, and 2.73±0.41, the cartilage damage was 2.07±0.53, 1.86±0.36, and 1.97±0.29, and the bone erosion was 2.33±0.68, 1.83±0.38, and 2.05±0.45, respectively, indicating that the histological findings were similar to each other (Fig. 2-A, B). TRAP staining was done to assess osteoclasts, playing a key role in bone erosion in rheumatoid arthritis. The results showed no difference in differentiation and total number of osteoclasts (Fig. 2-A, C). Similar to the macroscopic progressive pattern of the arthritis, the histological findings did not show improvement or exacerbation of the arthritis due to HPE in terms of synovial inflammation, cartilage damage and bone erosion.

At the end of the experiment we measured ,in the serum of mice, the concentrations of IL-1β, TNF-α, and IL-6, which are important inflammatory cytokines involved in the occurrence of rheumatoid arthritis, IL-10, which represses chronic inflammation in rheumatoid arthritis, and sRANKL, which is fundamental to the differentiation and activation of osteoclasts. IL-1β and TNF-α were also measured at the ankle joints. Serum concentrations of IL-1β in the respective groups were 235.38±15.11 pg/ml, 221.35±19.35 pg/ml, and 224.03±18.22 pg/ml. IL-10 concentrations were 204.86±36.20 pg/ml, 186.51±22.36 pg/ml, and 212.83±55.06 pg/ml. TNF-α concentrations were 659.60±249.53 pg/ml, 559.30±151.46 pg/ml, and 678.45±278.31 pg/ml. IL-6 levels were 566.90±127.04 pg/ml, 568.80±52.98 pg/ml, and 568.98±144.31 pg/ml, while sRANKL concentrations were was 1,472.62±174.03 pg/ml, 1,323.32±161.28 pg/ml, and 1,410.87±367.16 pg/ml. This indicated that there were no decreases in IL-1β, TNF-α, IL-6, and sRANKL levels and no increases in IL-10 level in groups exposed to HPE. Exacerbation of inflammatory cytokines was also not found (Fig. 3).

The severity of arthritis was assessed in the CIA model where autoimmune arthritis was induced. The arthritis scores in the vehicle and HPE administration groups were 3.75±2.22 and 4.00±1.83 on the 10th day and 6.04±1.91 and 5.75±1.25 at the end of the experiment, respectively. Similar arthritis severity findings were found for the two groups (Fig. 4-A). The ankle thickness in the respective groups was 2.94±0.18 mm and 2.91±0.11 mm at the beginning of the experiment, 3.31±0.34 mm and 3.58±0.55 mm on the 10th day, and 4.06±0.53 mm and 4.02±0.42 mm at the end of the experiment (Fig. 4-B). No significant difference was found in improvement or exacerbation by HPE between the two groups. Radiography using digital mammography and histopathologic assessment based on HE staining, which were done to assess the effect of HPE on destruction of bones (which is an important treatment goal in rheumatoid arthritis), clearly showed destruction of bone in both control and experimental groups (Fig. 5-A). The mean radiological score of the two groups was 2.67±0.34 and 2.82±0.25, respectively, indicating that bone destruction was mild in both groups (Fig. 5-B). The histological analysis scores in the respective groups were 3.06±0.20 and 3.22±0.36 for synovial inflammation, 2.27±0.33 and 2.32±0.41 for cartilage damage, and 2.29±0.28 and 2.48±0.31 for bone erosion, showing no improvement or exacerbation by HPE (Fig. 5-A, C).