The effect of dexamethasone on tissue water distribution and proton relaxation in Panc02 tumors

Magnerx Resonance Imaging. Vol. 5, pp. 483-492. Printed in rhs USA. All rightr rewvrd.

1987

Copyright 0

0730-725X/87 $3.00 + .OO 1987 Pergamon Journals Ltd.

l Original Contribution

THE EFFECT OF DEXAMETHASONE ON TISSUE WATER DISTRIBUTION AND PROTON RELAXATION IN PANC02 TUMORS PAUL G. BRAUNSCHWEIGER,*

KENNETH REYNOLDS,*

THOMAS R. NELSON,?

AND ELLEN MARING* *Department

of Experimental Therapeutics, AMC Cancer Research Center, 1600 Pierce Street, Denver, Colorado 80214, TDepartment of Radiology, University of California, San Diego, San Diego, California 92093

The present experiments were conducted to determine the effects of dexamethasone mediated changes in tumor water distribution on proton relaxation times (Tl, T2) in a murine pancreatic adenocarcinoma (PancO2). Spin lattice (Tl) and spin-spin (T2) relaxation times were determined by ex vivo methods (10MHz) and by in vivo imaging techniques (6.25 MHz) at various intervals after single or multiple dexamethasone treatments. In complementary studies, dexametbasone mediated changes in tumor capillary permeability, tumor water distribution, relative tumor blood flow and tumor cell proliferation were also determined. Proton spin lattice (2’1) and spin-spin (T2 relaxation times for PancO2 tumors shortened within two hours of a single dexamethasone treatment. The time course and magnitude of this response was dexamethasone dose dependent. The time dependent changes in Tl and T2 after dexamethasone were similar at 10 MHz (ex vivo) and 6.25 MHz (in vivo imaging). Although dexamethasone produced little or no change in total tumor water content and tumor cell proliferation, transient changes in the physiologic distribution of tumor water were clearly demonstrated. The data supports the idea that dexamethasone induced changes in the distribution of tumor water were mediated by changes in capillary permeability and tumor blood flow. These physiologic responses produced serial changes in tumor extracellular extravascular water content that were consistent with the observed changes in tumor Tl and T2. The results from these experiments might imply that therapy associated changes in tumor proton relaxation times may not only reflect changes in tissue water content, but may also reflect physiologic responses which alter the distribution of tissue water and solute. Keywords:

Proton relaxation:

Tumors; Vascular function;

INTRODUCTION

sues, however, the situation is more complex as observed proton relaxation rates reflect both cellular and extracellular water components. Tissue water content and the distribution of water and solutes which promote proton relaxation are subject to regulation by physiologic mechanisms which regulate capillary permeability, lymphatic drainage and blood flow. In relation to normal tissues, solid tumors characteristically demonstrate increased capillary permeability 14,23 low blood flow ‘8*29 poor lymphatic drainage and increases in extradellular water.‘2,14*24,28 Studies with norma11,1’,2’ and tumor tissues12*‘3 have

The physiologic significance of changes in tissue proton relaxation rates during the development of solid tumors and subsequent to therapy is not well understood. This reflects the complexity of biological responses in tissues during tumor development and therapy and the multiplicity of variables that influence tissue proton relaxation rates. At the cellular level, morphologic4s25s2” and biophysica12~‘,‘9~33 data suggests that most, if not all, cell water is ordered and exhibits some degree of motional restriction.16 In tisRECEIVED4/29/87;

Dexamethasone.

ACCEPTED 8/28/87.

was supported by Ca 39596 awarded by DHEW, and by a gift to AMC Cancer Research Center from Steven W. Farber. Address correspondence to Paul G. Braunschweiger, PhD.

authors would like to acknowledge the technical assistance of Kitty Orchard and Janice Renicker during the in vivo NMR studies, and thank Ms. Kara Selby for preparation of the manuscript. This work Acknowledgments-The

483

484

Magnetic Resonance Imaging 0 Volume 5, Number 6, 1987

suggested that physiologic responses which prompt changes in tissue water and solute distribution can profoundly effect tissue proton relaxation rates. In murine solid tumor models, dexamethasone treatments can temporarily change the physiologic distribution of tumor tissue water through transient, nontoxic perturbations of tumor capillary permeability. I4 Since dexamethasone can also produce transient changes in tumor blood flow and, in some tumors, cell proliferation,9~14 this non-cytoxic, hormonal agent has been employed in the present experiments to extend our previous studies1’,‘3,15 on the relationships between physiologic responses to hormones and changes in tissue proton relaxation. METHODS

Animals and Tumor Models A murine pancreatic adenocarcinoma (Panc02), originally developed by Corbett, et a1.,17 was obtained from the NC1 tumor bank, and maintained in 4-6 week old C57B1/6J male mice (Jackson Laboratories, Bar Harbor, ME) by subcutaneous implantation of tumor fragments on the right flank every 2 weeks. The volume doubling time for this tumor is approximately 3 days and mice with spherical tumors were initiated into study, 14-16 days after passage, when tumor diameters were approximately 1.4 cm (0.7 g). All mice were housed lo-12 per cage in a temperature and humidity controlled facility with a 12 hour lightdark cycle. All mice were fed standard mouse chow (Purina, Evansville, Indiana) and water ad libitum. Dexamethasone (Dexamethasone Phosphate, ES1 Pharmaceuticals, Cherry Hill, NJ) was freshly prepared prior to use and administered by intraperitoneal injection in 0.2 ml. Dexamethasone was administered as either a single injection (5,lO mg/kg) or as 10 mg/kg given every 12 hours for 3 doses.

Ex Vivo IH-NA4R Measurements Proton relaxation times (Tl,

T2) for excised tumor tissues (ex vivo) were determined with a bench top proton spin analyzer (RADX Corporation, Houston, TX) operating at 10 MHz as previously described.“~“~r3 Animals were killed by cervical luxation and the tissues quickly excised and weighed. The tissues were minced with scissors and the tumor fragments packed into 5 mm NMR tubes. These preparations were then centrifuged for 5 minutes (1500 rpm), at room temperature, to remove air spaces. The samples were then heated to 37°C in a block heater and relaxation times (Tl, T2) immediately read. All samples were read within 1 hour of excision and the mean

of triplicate spin analyzer readings (a total of 24 Tl and T2 determinations) was determined for each sample. Tl and T2 relaxation times were corrected for machine variation using daily calibration curves obtained with the commercial Tl and T2 standards provided by the manufacturer (RADX Corporation, Houston, TX).

In Vivo IH-NMR

Measurements

Proton NMR relaxation times for Panc02 tumors were also determined in vivo, by imaging techniques, with a 0.15 Tesla (6.25 MHz) wholebody clinical MRI unit (Technicare Corporation). Tumor bearing mice were anesthetized with 0.2 ml of a Ketamine (70 mg/kg), Xylazine (3 mg/kg) and acepromazine (4 mg/kg) cocktail by intraperitoneal injection. Anesthesia, lasting at least 60 minutes, was obtained within 3-4 minutes. Animals were placed in a Styrofoam jig in the homogeneous region of a 6-inch receiver coil. Copper sulfate standards in 2.5 cm (ID) sealed glass vials were simultaneously imaged to correct for machine variation. The slice thickness was 0.75 cm and tumor localization was obtained from multi-slice scout images. Anesthesia and set up time was usually about 15 minutes. The 256 x 256 images were constructed by utilizing a two dimensional Fourier transform (2DFT) phase encoding acquisition algorithm with a standardized spin-echo technique (SE) and modified Carr-PurcellMeiboom-Gill (CPMG) multi-echo sequence. The signal intensity (I) for the nth echo in the CPMG pulse sequence is defined by: I=

~N(H)~-“‘*TE/T~

[ 1 _ +

2e-(TR+TE/2-n*TE)/TI e-TR/71

1

(1)

where TR is the repetition time, TE is the interval between the 90” pulse and the echo peak, n is the number of ethos acquired per pulse, n’ is the echo number, k is a proportionality constant and N(H) is the proton density. The average signal intensity for the region of interest (ROI) in images obtained with a fixed TE (30 msec) and TR’s of 63, 125, 250, 500, 1000, 2000 and 4000 msec was used to calculate the spin-lattice relaxation time (Tl) by fitting the data to Eq. (1) with a chi square error minimization model.7 A constant percentage of signal intensity reduction due to T2 decay was assumed independent of TR. The average signal intensity in the ROI for TE’s of 30, 60, 90, 120, 150, 180, 210, and 240 msec with a TR fixed at 2000 msec was used to compute the spin-

Dexamethasone mediated changes in proton NMR Tl, T2 0 P. G.

spin (T2) relaxation time using Eq. (1) and the chi square fitting method. Equation (1) assumes that perfect 90 and 180 degree pulses were used in the experiments. This condition is rarely met with clinical imagers and contributes to the error in making T 1 and T2 measurements. In these in vivo experiments Tl and T2 obtained for tumor tissues were corrected for daily machine variation using simultaneously imaged CuS04 standards and least squares linear regression analysis.

Tumor Water Distribution and Vascular Permeability Total tumor water content was measured by gravimetric methods as previously described.‘2J4 Tumor tissue was immediately resected, placed in preweighed glass vials, reweighed and heated at 105°C until a constant weight was achieved. The difference in tissue mass before and after drying was taken as the total water content. After correction for ambient conditions, the water mass was expressed as ~1 of water/ gram tissue. Tumor plasma and extracellular water distribution was determined by 1251-BSA and “Cr-EDTA isotope dilution techniques, described previously. 12,14*23,24 Approximately 300,000 cpm 1251-BSA and 400,000 mixed in 0.1 ml was injected via cpm 51Cr-EDTA, the lateral tail vein. At 1, 5, 10, 20, 30 and 60 minutes, venous blood samples were obtained, in cohorts of 5 mice, and the mice immediately killed by cervical luxation. The subcutaneous tumors were rapidly reflected, frozen in situ (dry ice, ethanol bath) and resected in toto. The tissues were weighed, and counted for radioactivity (“Cr, 1251) in a 3 channel, gamma well spectrometer (Packard Instrument Co. Downer’s Grove, IL). The blood samples were centrifuged and the plasma counted for “‘1 and 51Cr radioactivity. Injection standards were prepared and counted with tissue samples. Radioactivity in tissue and plasma was expressed as a percent of the injected dose per gram tissue on per ml plasma. The dilution volumes for the labeled probes were calculated from: %ID/g

tissue + %ID/pl

plasma

= pi/g tissue

(2)

Since tumor capillaries are more permeable to serum proteins than normal tissue capillaries, tumor plasma volumes were estimated by extrapolating pi/g (Iz51BSA) vs. time to t=O. I4 The Y intercept, its 95% confidence interval, the slope and its 95% confidence interval were determined by linear least squares regression analysis. 12,r4 Extracellular water volumes were estimated when

BRAUNSCHWEIGER ET AL.

485

serum and tissue 51Cr-EDTA were in equilibrium. That is, when the 51Cr-EDTA pi/g reached plateau levels. ‘2x14,24Plateau values were taken as the asymptote for the line of best fit with the equation Y = A-Be’“. The line of best fit was determined by an iterative least squares computer curve fitting routine. Plateau values were usually observed by 20 minutes purposes, the after injection. i2,i4 For comparative standard deviation for the “Cr-EDTA extracellular distribution volume was calculated from the values obtained for the 10 tumors analyzed at the 20 and 30 minute study intervals.

Tumor Blood Flow 86RbC1 distribution in tumors, 45 set after intravenous injection (lateral tail vein) was used to assess relative tumor blood flo~.‘~,~’ Ten tumors per study interval were employed and the % ID/g tumor was taken as the distribution of the cardiac output per the relative gram tumor. 3o In some experiments, blood flow was also determined in femoralis muscle as above. Cell Proliferation The effects of dexamethasone on tumor cell proliferation were assessed by in vivo 3H-Thymidine labeling and autoradiography. Tumor bearing mice were administered 3H-Thymidine (0.5 &gram body weight; 20 Ci/mMole; NEN, Boston, MA) in 0.2 ml 0.9% NaCl by intraperitoneal injection. Mice were killed 30 minutes later, the tumors minced, and the cells dissociated with a 45 minute enzyme (type III tryspin, 0.75 mg/ml; type II collagenase, 0.75 mg/ml; crude DNAase, 0.05 mg/ml; Sigma Chemical Company, St. Louis, MO) treatment (20 ml enzyme cocktail per 0.3 grams tissue) at room temperature. The cells were washed with McCoy’s media + 10% fetal calf serum and applied to microscope slides with a cytocentrifuge. Autoradiograms were prepared using liquid photographic emulsion (NTB2; Kodak, Rochester, NY) as previously described. I4 Two week exposure times, determined by test sets, provided a mean grain density in excess of 50 grains/labeled nucleus. Background was usually less than one grain per equivalent cell area and cells with 3 or more grains per nucleus were considered labeled.

Statistical Analysis Group means and the standard deviation of the mean were calculated by conventional methods. Students t test was used to assess the significance of dif-

486

Magnetic

Resonance

Imaging

0 Volume

ferences between means and P I 0.05 was considered adequate justification to reject the null hypothesis.

5, Number 6, 1987

pare proton relaxation times for Panc02 tumors in live, anesthetized mice at 6.25 MHz and for excised tumor tissue (10 MHz) after dexamethasone treatment. In these experiments, tumor bearing mice were anesthetized, imaged at 6.25 MHz, allowed to recover from the anesthesia, killed, and tumor tissue prepared for ex vivo analysis. In vivo at 6.25 MHz, the Tl for 54 control tumors was significantly shorter than the Tl’s seen by ex vivo assay at 10 MHz. The intertumor coefficient of variation (CV) was 9% for in vivo measurements and 6% for ex vivo measurements. The intratumor CV for the in vivo Tl measurement in individual tumors ranged from 6% to 23% for individual tumors with a mean of 10.6% (8.8, 12.4; 95% confidence interval). The CV for simultaneously imaged, 2.5 cm CuSO4 standards was 6.1%. The T2’s

RESULTS Figure 1 shows the results from experiments to assess the time and dose dependent effects of a single dexamethasone treatment on proton relaxation times (10 MHz) for subcutaneous Panc02 tumors. In these experiments, 10 mg/kg produced a 30% and 45% decrease in Tl and T2 times, respectively, at 10 hours after treatment. At 6 hours, after 5 mg/kg, Tl and T2 times were decreased by about 17%. Recovery was initiated by 8 hours after 5 mg/kg and by 12 hours after 10 mg/kg. Table 1 shows the results from experiments to com-

A

B

700

--a

--a 5mg/kg -lOmg/kg

Smg/kg

--rlOmg/kg I

I

I

4

8

12

I

.

16 20

I

24

4

6

12

16

20

24

Hrs

Hrs

Fig. 1. The effect of a single intraperitoneal dose of dexamethasone (0, 5 mg/kg; H, 10 mg/kg) on the Tl (1A) and T2 (1B) times (10 MHz) for Panc02 tumors. Each point represents the mean & 1 SEM for 5 tumors.

ex vivo proton relaxation

Table

1. The effect

of a single 10 mg/kg Dexamethasone treatment for Panc02 tumors at 6.25 and 10 MHz

on Tl and T2

Tl

T2

6.25 MHz Controls

t=o

t=4

t=L3

495 460 385 376

* k t -t

44 13 13 13

(54)* (5) (5)t (5)t

10 MHz 695 600 518 456

*msec; mean -t 1 SD for (n) tumors. tsignificantly (P < 0.05) different from t = 0 controls.

+ + + f

42 17 10 13

(54) (5) (5)T (5)t

6.25 MHz 118 117 94 67

f 15 (54) * 6 (5) + 2 (5)t +- 4 (5)-t

10 MHz 125 115 98 89

? * f *

14 (54) 3 (5) 3 (5)t 4 (5)t

Dexamethasone mediated changes in proton NMR

Tl.

T2 0 P. G.

BRAUNSCHWEIGER ET AL.

487

anesthetized and unanesthetized mice by 8 hours, it is unlikely that changes in the distribution of the cardiac output to the tumor, as a result of the 15 minute anesthesia, affected proton relaxation parameters. Table 3 shows the results from serial imaging experiments in which proton relaxation times were determined for individual Panc02 tumors at 4 and 8 hours after 10 mg/kg dexamethasone. In these experiments pretreatment scans were obtained for 3 groups of 5 tumors, 24 hours prior to dexamethasone treatment. Second scans were then obtained prior to (t=O controls), at 4 hours, or at 8 hours after dexamethasone. Tumor Tl values were significantly (PcO.05) shorter than respective pretreatment values at 4 and 8 hrs after dexamethasone. In this experiment, T2 was shorter only at the 8 hour study interval. These results suggest that the changes in proton relaxation observed by in vivo imaging were probably not related to anesthesia mediated changes in tumor blood flow, but to dexamethasone mediated responses effecting water and solute distribution. To test this hypothesis further, the time dependent changes in proton NMR relaxation times for Panc02 tumors were determined after 3, 10 mg/kg dexamethasone treatments given 12 hours apart. The results from replicate ex vivo studies (10 MHz) and a single in vivo imaging study (6.25 MHz) are shown in Figs. 2 and 3. The time course for the dexamethasone induced change in tumor T 1 was reproducible in 2 ex vivo experiments. In tumors where Tl was first determined in vivo (6.25 MHz) and then ex vivo (10 MHz),

determined for control tumors at 6.25 and 10 MHz were not significantly different. In the experiment to assess the effect of dexamethasone on Panc02 proton relaxation times, cohorts of 5 mice were studied at 0, 4 and 8 hours after a single 10 mg/kg treatment. Although Tl’s at 6.25 MHz were significantly shorter than those seen at 10 MHz, the relative change in Tl was similar at both field strengths. Similarly, the T2 responses seen at 10 and 6.25 MHz were not significantly different. Imaging experiments were of necessity conducted in anesthetized mice and we had some concern that the physiologic consequences of cardiosuppression could effect observed proton relaxation in the tumor. Table 2 shows the results from an experiment, conducted at 10 MHz, to assess the effect of Ketamine, Acepromazine, and Xylazine anesthesia on tissue blood flow and proton relaxation in dexamethasone treated mice. The anesthesia cocktail used in the imaging experiments clearly perturbed tumor and muscle perfusion as determined by 86RbCl distribution. In anesthetized mice, the tumor %ID/g was significantly increased (55+ lo%, mean & 1 SD) as compared to that for tumors from unanesthetized controls. Conversely, the anesthesia significantly decreased 86RbCl distribution in femoralis muscle by 53+8%. In unanesthetized mice, dexamethasone significantly decreased vascular perfusion in both Panc02 tumors and in femoralis muscle. In anesthetized mice, no such effect was observed up to 8 hours. Since dexamethasone resulted in similar tumor Tl and T2 shortening in

Table 2. The effect of anesthesia on proton relaxation and tumor blood flow in Panc02 tumors after a single 10 mg/kg

Dexamethasone

treatment

Anesthesia*

No Anesthesia

Control

Tumor

Muscle

Tumor

Muscle

%ID/g$ Tl§ T2

3.81 L 0.89 678 ? 18 136+3

1.22 + 0.38 487 + 27 54 * 3

2.61 * 0.22 679 k 16 136& 3

3.23 ? 0.38 473 * 11 49 + 3

4 hrs VOID/g Tl T2

3.44 * 1.31 610 * 13 116?3

1.32 + 0.31 470 + 29 47 * 4

2.23 k 0.36 595 * 14 117+-3

2.71 ? 0.71 471 + 13 51+3

3.75 + 1.02 578 * 18 108 f 9

1.31 f 0.33 430 f 16 45 * 13

2.26 i 0.47 580 * 25 102 + 4

2.42 ? 0.47 420+31 43* 11

8 hrs % I D/g

Tl T2

*Ketamine, 70 mg/kg; Xylazine, 3 mg/kg; Acepromazine, 4 mg/kg given in 0.2 ml, ip, 15 minutes tmean+l SDn=5. tg6RbCl distribution per gram tissue as a % of the injected dose (ID). §Tl and T2 relaxation times in msec.

prior to tissue sampling.

Magnetic Resonance Imaging 0 Volume 5, Number 6, 1987

488

Table 3. The effect

Dexamethasone on observed proton relaxation times for PancO2 tumors in vivo at 6.25 MHz

of 10 mg/kg

24 hrs Pretreatment Control T2

488 +- 13* (5) 119 k 7 (5)

Tl T2

507 l?r 8 (5) 11627 (5)

-

Tl T2

550 + 15 (5) 118 ? 5 (5)

-

Tl

t=4

f=8

-

-

446 + 12 (5)t 115 * 6 (5)

-

t=o

503 * 11 (5) 121 + 4 (5)

*mean i I SD (n). tsignificantly different

from pretreatment,

457 * 11 (5)t 103 +- 4 (5)

P < 0.05.

15 700

14

f

600

121

% g

500

I=

400

o

81 0,

o

1

@,A

6.25 MH. 10.0

6.25Ml-t~

A 10.0

MH2

71

MH:

300 0

12

24 Hrs after

36

46

60

Dex

0

12

24

48

36

Hrs after

60

Dex

Fig. 2. The effect of three 10 mg/kg dexamethasone treatments on the Tl relaxation times for PancO2 tumors in vivo (0; 6.25 MHz) and in 2 replicate ex vivo experiments (0, A, 10.0 MHz). Tumors studied in vivo (0) were subsequently excised and studied at 10 MHz (0). Each point is the mean * 1 SEM for 5 tumors.

Fig. 3. The effect of three 10 mg/kg dexamethasone treatments on the T2 relaxation times for PancO2 tumors in vivo (0; 6.25 MHz) and in 2 replicate ex vivo experiments (0, A, 10.0 MHz). Tumors studied in vivo (0) were subsequently excised and studied at 10 MHz (0). Each point is the mean k 1 SEM for 5 tumors.

similar responses to dexamethasone treatment were seen. Recovery of tumor Tl was initiated by 12 hours after the treatment and pretreatment T l’s were seen by 24 hours. The T2 response (Fig. 3) was somewhat more variable. However, the in vivo and ex vivo

responses were temporally similar. In Panc02 tumors, the TI and T2 response after dexamethasone treatments had similar time courses. Since previous studies with RIF-1 and Panc02 tumors indicated a close relationship between changes

Dexamethasone

mediated

changes

in proton

in tissue water distribution and changes in proton relaxation after dexamethasone treatments,12 parallel studies were conducted to measure the time dependent changes in tumor water distribution, vascular permeability and blood flow (Table 4 and Fig. 4). Total water content was significantly decreased within 4 hours of a single 10 mg/kg dexamethasone treatment (Table 4). Between 4 and 24 hours after 3, 10 mg/kg treatments total water content was reduced by about 60 pi/g (70/o). Total water content approached pretreatment values by 48 hours. Tumor plasma volumes were significantly increased within 2 hours after a single dexamethasone treatment. This increase in plasma volume was associated with a marked decrease in capillary permeability and dehydration of the total extracellular space. By six hours after treatment, extravascular, extracellular water (ECV-PV) was decreased by about 40%. Between 4 and 24 hours after dexamethasone, plasma volumes were increased by approximately 60% (P < 0.02). Capillary permeability also increased during this interval and at 24 hrs, the ‘*jI-BSA extravasation rate was significantly greater than that in untreated tumors. In Fig. 4, it can be seen that this marked increase in capillary permeability was accompanied by a marked increase in tumor interstitial water content. Tumor blood flow, as indicated by *‘jRbCl distribution, was unchanged at 4 hours, but significant increases were seen at 12 and 24 hours after treatment (Table 4). Extravascular, extracellular water (ECVPV) increased from about 280 PI/g at 4 hours to about 530 PI/g by 24 hours after dexamethasone (Fig. 4).

Table

4. The effect

of Dexamethasone Total Water cLl/g

Control 10 mg/kg 2 hrs 4 hrs 6 hrs

(x l)$

10 mh/kg 4 hrs 12 hrs 24 hrs 36 hrs 48 hrs 12 hrs

(x3)$

NMR

j-1, T2 0 P. G. BRAUNSCHWEICER

DISCUSSION The content and physiologic distribution of water in normal and tumor tissue can be profoundly influenced by corticosteroid hormones.‘2.‘4,3’ In corticosteroid receptor containing RIF-I and Panc02 tumors, we previously showed that 3, 10 mg/kg treatments, given 12 hours apart led to decreases in tissue interstitial water content and increases in plasma water content secondary to dexamethasone mediated changes in vascular permeability.12~‘4 Although Panc02 tumors exhibit lower levels of specific dexamethasone receptors than do RIF-1 tumors,12 water mobilization after dexamethasone was more profound than in the RIF-1 model.” Vascular permeability decreased within 2 hours of a single 10 mg/kg treatment and by 6 hours the extravasation of ‘251-BSA was reduced by nearly 75% in the Panc02 tumors. Further, total tumor water and extracellular water dis-

and vascular

ECV 0g

functions

in Panc02

Permeability pl/g/min

841 + 11*

19 f 3-t

395 + 1s*

1.43 * 0.07t

824 f 4 790 + I 799 ?z 13

26 i 2 25 f 3 24 i- 2

352 f 12 296 * 14 250 * 10

0.80 i 0.03 0.43 f 0.01 0.36 + 0.02

780 780 780 810 820 830

33 30 29 21 22 24

310? 18 300 k 28 560 + 98 310 + 15 380 + 26 415 AZ57

0.70 0.98 2.08 1.23 1.26 1.41

? * k * * *

6 9 7 10 12 8

*Mean t 1 SD for 10 tumors. tMean ? 95% confidence interval. $x I, one treatment; x3, 3 treatments.

*4 * 4 + 6 t4 + 3 + 3

489

Capillary permeability, tumor extracellular water distribution and tumor blood flow decreased between 24 and 36 hrs after treatment. Pretreatment tumor water distribution, capillary permeability and tumor blood flow were re-established by 48 hours after treatment. Also shown in Fig. 4 is the effect of dexamethasone on the 3H-Thymidine labeling index for Panc02 tumors. Although dexamethasone produced profound vascular responses that were temporally coincident with changes in proton relaxation, this hormone had little or no effect on cell proliferation in Panc02 tumors.

on water distribution Plasma Volume pi/g

ET AL.

+ 0.12 + 0.07 ? 0.05 k 0.13 +- 0.07 + 0.07

tumors

86RbCl ‘?“o ID/g 3.79 * 0.49*

-

3.65 5.13 5.41 2.96 3.53 4.41

+ + + f * *

0.66 0.81 0.59 0.42 0.39 0.31

Magnetic Resonance Imaging 0 Volume 5,

490

4

50

3-

40

3

30

z

20

c 0 I m

10 0

n i \

400

i 2

300

200

D

D

24

D Hrs

48 after

72

96

Dex

Fig. 4. The effect of three 10 mg/kg dexamethasone treatments on the 3H-Thymidine labeling index (‘H-dThd LI, A, mean k 1 SEM for 5 tumors), interstitial water content (m, ISW = ECV-PV) and capillary permeability (0, means from Table 4) in PancO2 tumors.

tribution were reduced by 5 and 35% respectively. Within 12 hours after 3 dexamethasone treatments, a vascular rebound was observed. At 24 hours, marked increases in “‘1-BSA extravasation, tumor edema, and vascular perfusion characterized this response. In contrast to our observations in other tumor models,” dexamethasone had little or no discernible effect on cell proliferation in Panc02 tumors. Marked changes in Panc02 tumor proton relaxation times were observed in dexamethasone treated mice by imaging techniques (6.25 MHz) and with excised tumor tissue (10 MHz). In most experiments, Tl values at 6.25 MHz were about 25% shorter than those at 10 MHz. These observations are consistent with the known field dependence for Tl and conforms to the relationship described by Bottomly.’ As previously observed in other systems,* the T2 for Panc02 tumors was not dependent on magnetic field strength. The relative changes in tumor proton relaxation times determined in imaging experiments with Live, anesthetized mice up to 8 hours after dexamethasone

Number

6, 1987

were similar to the changes seen from ex vivo measurements at 10 MHz. Further, the time dependent response during vascular rebound seen with excised tissues was not unlike that observed in vivo. While it may be argued that sample preparation may have altered the tumor tissue architecture, the changes in tissue proton relaxation, seen ex vivo, seemed to provide a reasonably good indication of changes observable by imaging techniques in live animals. Few studies have addressed this question directly, but this conclusion is not inconsistent with the analysis of previously published data for unperturbed normal tissues.’ Preanesthesia with Ketamine, Acepromazine and Xylazine had no effect on the measured Tl or T2 (37°C) for excised Panc02 tissue even though the 15 minute anesthesia produced acute changes in the distribution of the cardiac output to the tumor and muscle. Since anesthesia undoubtedly reduced the cardiac would output, an increase in the 86RbCI distribution reflect compensatory mechanisms to maintain a constant tumor blood flow. Decreases in ‘“RbCI uptake in muscle, however, may reflect an anesthesia related decrease in blood flow to tQe extremities.34 Anesthesia related changes in body temperature could have influenced the in vivo Tl and T2;2’ however, such influences are probably small and at least consistent in all in vivo experiments. Although dexamethasone resulted in slight PancO2 tumor dehydration (7%), the attendant shortening of Tl and T2 was quantitatively related to decreases in tumor extracellular water. During the first 6 hours after dexamethasone, strong linear relationships (r’c0.95) were noted between ECV and Tl and ECV and T2. The recovery time course for proton relaxation was found to be similar to that for changes in capillary permeability and tumor water distribution in dexamethasone treated tumors. Although neither of these parameters would directly affect proton relaxation, such physiologic responses could clearly have indirect effects. We might speculate that dexamethasone mediated decreases in capillary permeability created new osmotic gradients between the plasma and interstitial space, promoted the diffusion of mobile water from the interstitium and resulted in the concentration of extracellular solute. As studies with model serum protein solutions have shown that proton relaxation times are highly dependent on solute quantitatively similar changes in concentration,‘2.‘3 proton relaxation and tumor ECV after dexamethasone might be explained by such a mechanism. At 24 hours after dexamethasone, the total tumor water content was similar to that at 4 hours, but the ECV was 40% greater than that for untreated controls.

Dexamethasone

mediated changes in proton NMR

Proton relaxation times at 24 hours were prolonged relative to the 4 hour groups but they did not exceed pretreatment values. Although seemingly paradoxical, this observation is not inconsistent with the physiowas logic data. At 24 hours, ‘251-BSA extravasation increased 45% relative to that in controls. Even though the ECV was markedly increased, the increases in capillary permeability would also promote an increase in extracellular solute content. Thus, while increases in the ECV would predictably increase relaxation times, an increase in solute content of the ECV would tend to have a negative influence on the magnitude of this response. Although our data implies that observed changes in tissue Tl and T2 after dexamethasone were prompted by vascular responses that perturbed extracellular water and solute concentrations, observed tissue proton relaxation times reflect a weighted summation of celluIar and extracelluIar relaxation rates. Corticosteroids have been shown to affect cell proliferation,23 and membrane fluidity2’ in experimental systems and such responses might influence proton relaxation rates. Proliferation related influences on cellular proton relaxation timesI may be trivial in dexamethasone treated Panc02 tumors, as dexamethasone had no effect on 3H-Thymidine labeling indices. The effects of dexamethasone on membrane fluidity in this model, however, remain to be determined. Corticosteroids may also stimulate protein synthesis in some cells.27 Although not studied here, such a response and the attendant influences on cellular water mobility can not be totally excluded. Nonetheless, the present studies in Panc02 tumors, previous studies in RIF-1 tumors12,‘3 and studies with rat ventral prostate” indicate that nontoxic hormone mediated changes in vascuIar permeability, tissue water distribution, and perhaps cell proliferation’*‘5 can markedly alter observed proton relaxation times for tumor tissues. Experiments in several tumor lines have indicated that cytotoxic therapies may also profoundly affect tumor blood flow, vascular permeability and tissue water distribution. 13s3*These studies implied that changes in tissue water distribution may reflect cell kill and, subsequently, the proliferative recovery of surviving tumor cells.13 Since the vascular responses, which in the present study were shown to influence tissue proton relaxation, may also influence therapeutically important parameters such as cell proliferation, tissue oxygenation and the distribution of systematically administered therapeutic agents, serial in vivo proton NMR measurements may provide a non-invasive way to assess initial response to therapy and to subsequentIy identify efficacious treatment intervals.

Tl,

T2 0 P. G. BRAUNSCHWEIGER ET AL.

491

REFERENCES 1. Bakay, L.; Kurland, R.M.; Parrish, R.G.; Lee, J.C.; Peng, R.J.; Bartkowski, H.M. Nuclear magnetic resonance studies in normal and edematous brain tissue.

Exp. Brain Res. 23:241-248; 1915. 2. Beall, P.T. States of water in biological systems. Cryobiology. 20:324-334; 1983. 3. Beall, P.T.; Asch, B.B.; Chang, D.C.; Medina, D.; Hazlewood, C.F. Distinction of normal, preneoplastic, and neoplastic mouse mammary primary cell cultures by water nuclear magnetic resonance relaxation times. J. Nat. Cancer Inst. 64:335-338; 1980. 4. Beall, P.T.; Brinkley, B.R.; Chang, D.C.; Hazlewood, C.F. Microtubule complexes correlated with growth rate and water proton relaxation times in human breast cancer cells. Cancer Res. 42:4124-4130; 1982. 5. Beall, P.T.; Hazelwood, C.F.; Chang, D.C. Microtubule organization and the self diffusion coefficient of water in baby hamster kidney cells as a function of temperature. J. Cell Biology, Vol. 95, p. 334a; 1982. 6. Beall, P.T.; Hazlewood, CF.; Rao, P.N. Nuclear magnetic resonance patterns of intracellular water as a function of HeLa cell cycle. Science 192:904-907; 1976. 7. Bevington, P.R. Data reduction and error analysis for the physical sciences, pp. 219-222, New York: McGraw Hill, 1981. 8. Bottomley, P.A.; Fogler, T.H.; Argersinger, R.E.; Pfeifer, L.M. A review of normal tissue hydrogen NMR relaxation mechanisms from l-100 MHz: Dependence on tissue type, NMR frequency, temperature, species, excision, and age. Med. Phys. 11:425-448; 1984. 9. Braunschweiger, P.G.; Ting, H.L.; Schiffer, L.M. Receptor dependent antiproliferative effects of corticosteroids in RIF tumors and implications for sequential therapy. Cancer Res. 42:1686-1691; 1982. 10. Braunschweiger, P.G.; Ting, H.L. and Schiffer, L.M. The correlation between glucocorticoid receptor content and the antiproliferative effect of dexamethasone in experimental solid tumors. Cancer Res. 43:4757-4761; 1983. 11. Braunschweiger, P.G.; Glade, L.M.; Maring, E.M.; Machus, K.; Reynolds, K. Biological implications of Androgen dependent changes in Proton-NMR relaxation times in rat ventral prostate. The Prostate 9:283294; 1986. 12. Braunschweiger, P.G.; Schiffer, L.M.; Ting, H.L. Proliferative recovery (PR) in solid tumors after cyclophosphamide. Proceedings of the Cell Kinetics Society, Cell Tissue Kinetics. 16:610, 1983. 13. Braunschweiger, P.G.; Schiffer, L.M.; Furmanski, P. The measurement of extracellular water volumes in tissues by Gadolinium modification of lH-NMR spin lattice (Tl) relaxation. Magn. Reson. Zmag. 4:285-291; 1986. 14. Braunschweiger, P.G.; Schiffer, L.M. Effect of dexamethasone on vascular function in RIF-1 tumors. Cancer Res. 46:3299-3303; 1986. 15 Braunschweiger, P.G.; Schiffer, L.M. IH-NMR relaxation times in solid tumors after dexamethasone. CeN Tissue Kinef. In press; 1986. 16. Clegg, J.S. Properties and metabolism of the aqeuous cytoplasm and its boundaries. Am. J. Physiol. 246: R133-R155: 1984.

Magnetic Resonance Imaging 0 Volume 5, Number 6, 1987

492

17. Corbett, T.H.; Robert, B.S.; Leopold, W.R.; Peckham, J.C.; Wilkoff, L.J.; Griswold, D.P.; Schabel, F.M. Induction and chemotherapeutic response of two transplantable ductal adenocarcinomas of the pancreas in C57B1/6 mice. Cancer Res. a44:717-726; 1984. 18. Gullino, P.M. Extracellular compartments of solid tumors. In: Cancer, A Comprehensive Treatise, Biology of Tumors: Cellular Biology and Growth, Becker, F.F. (ed.), Plenum Press, NY, pp. 327-354; 1975. 19. Hazelwood, C.F.; Nichols, B.L.; Chamberlain, N.F. Evidence for the existence of a minimum of two phases of ordered water in skeletal muscle. Nature, London 222~747-750; 1969. 20. Keating, K.M.; Roess, D.A.; Peacock, J.S.; Barisas, B.C. Glucocorticoid effects on membrane lipid mobility during differentiation of murine B lymphocytes.

Biochimica et Biophysics Acta. 846:305-312; 1985. 21. Lee, J.C.; Bakay, L. Ultrastructural changes in the edematous central nervous system. Arch. Neurol. 14: 36-49; 1966. 22. Nelson, T.R.; Sung, S.M. Temperature dependence of Proton relaxation times in vitro. Mugn. Reson. Imag. 5:189-199; 1987. 23. Peterson, H.-I.; Appelgren, L.; Lundborg, G.; Rosengren, B. Capillary permeability of two transplantable rat tumors as compared with various normal organs of the rat. VIIth Europ. Conf. Microcirculation, Bibl. Anat. (Basal) 12:511-518; 1973. 24. Peterson, H.-I.; Appelgren, L.; Kjartansson, I.; Selander, D. Vascular and extravascular spaces in a transplantable rat tumor after local x-ray irradiation. 2.

Krebsforsch 87: 17-25; 1976. 25. Porter, K.R.; Tucker, J.B. The ground substance of the living cell. Scientific American 244(3):56-67; 1981.

26. Porter, K.R.; Boggs, D.P.; Anderson, K.L. The distribution of water in the cytoplasm. Proc. 40th Annual EMSA meeting: 4-7; 1982. 27. Pratt, W.B.; Aronow, L. The effect of glucocorticoids on protein and nucleic acid synthesis in mouse fibroblasts growing in vitro. J. Biol. Chem. 241:5244-5250; 1966. 28. Reed, R.K.; Wiig, H. Interstitial albumin mass and transcapillary extravasation rate of albumin in DMBAinduced rat mammary tumors. Scan. J. C/in. Lab. Znvest. 43:503-512; 1983. 29. Reinhold, H.S. In vivo observations of tumor blood flow. In: Tumor Blood Circulation, Peterson, H.-I. (ed.), pp. 115-128, CRC Press, FL: Boca Raton; 1979. 30. Saperstein, L.A. Regional Blood Flow by fractional distribution of indicators. Am. J. Physiol. 193:161-168; 1958. 31. Sato, H.; Hashimoto, M.; Sugio, K.; Ohuchi, K.; Tsurufuki, S. Comparative study between steroidal and non steroidal anti-inflammatory drugs on the mode of their actions on vascular permeability in rat carrageeninair-pouch inflammation. J. Pharm. Dyn. 3:345-352,

1980. 32. Schiffer,

L.M.; Braunschweiger, P.G. Water compartmentalization, vascular function and cell proliferation in RIF-1 tumors after cyclophosphamide. Proc. AACR 26:43; 1985. 33. Seitz, P.K.; Chang, D.C.; Hazelwood, C.F.; Rorschach, H.E.; Clegg, J.S. The self-diffusion of water in Artemia cysts. Arch. Biochem. Biophys. 210:517-524;

1981. 34. Zanelli, G.D.; Lucas, P.B.; Fowler,

J.F. The effect of anesthesia on blood perfusion in transplantable mouse tumors. Brit. J. Cancer. 32:380-388; 1975.