.mmmmmnRmmmRmmnmRMMRRRRR
`
`organismal fitness (for example, metabolic
`(2, 17-21) and
`efficiency, growth rate)
`have concluded that allozyme polymor-
`phisms themselves underlie physiological
`energetic differences by virtue of their in-
`fluence on metabolite flux through central
`biochemical pathways (5, 22). Such bal-
`ancing selection on allozyme polymor-
`phisms could counter the influence of ge-
`netic drift, even in the face of population
`subdivision due to historical demographic
`events or contemporary restrictions on gene
`flow that are reflected in geographically
`divergent frequencies of neutral genetic
`markers. Hidden molecular variation that
`would also distinguish Atlantic and Gulf
`oysters may exist at the allozyme loci, but
`this possibility can in a sense be subsumed
`under the hypothesis of balancing selec-
`tion, provided that the selection operates
`with respect to the observed electromorph
`classes instead of the level of the hidden
`variation. A slower rate of evolution for
`allozymes as compared to scnRFLPs can also
`be eliminated, because both cases in oysters
`involve the population level sorting of an-
`affect both
`cestral polymorphisms that
`equally.
`Regardless of the specific underlying
`causes, the heterogeneity in geographic pat-
`scnDNA, and
`tern among allozyme,
`mtDNA data sets cannot be accommodated
`under a single evolutionary model involv-
`ing either neutrality or balancing selection
`(23). In this example where an allozyme
`survey had suggested high levels of gene
`flow, dramatic population genetic separa-
`tion nonetheless was present in both the
`nuclear and the cytoplasmic genomes. Re-
`sults of this study emphasize the need for
`caution in inferring population genetic
`structure and gene flow from any single class
`of genetic markers.
`
`REFERENCES AND NOTES
`1. N. E. Buroker, J. Fish. Res. Board Can. 36, 1313
`(1979).
`2. E. Zouros, S. M. Singh, H. E. Miles, Evolution 34,
`856 (1980).
`3. T. J. Hilbish and R. K. Koehn, ibid. 39, 1302
`(1985).
`4. _, Science 229, 52 (1985).
`5. R. K. Koehn, A. J. Zera, J. G. Hall, in Evolution of
`Genes and Proteins, M. Nei and R. K. Koehn, Eds.
`(Sinauer, Sunderland, MA, 1983), pp. 115-136.
`6. J. B. Mitton and M. C. Grant, Annu. Rev. Ecol.
`Syst. 15, 479 (1984).
`7. N. E. Buroker, Mar. Biol. 75, 99 (1983).
`8. C. A. Reeb and J. C. Avise, Genetics 124, 397
`(1990).
`9. J. C. Avise, Oikos, in press.
`10. A. C. Wilson et al., Biol. J. Linn. Soc. 26, 375
`(1985).
`11. Sequences of the primers 5' to 3' are as follows:
`CV-7: AAGCTTTAGCCTTCAACTCAGACAA,
`AAGCTTTAAGGTAGAAGCAAATTA; CV-1 9:
`TAAGTTG11TCTGATCCTTMG, CMATrT1M-
`TACCATTGCC; CV-32: GGAAGCTTTAT-
`TATCTAACAGTCT, GGAAGCTTACAAAACAAGC-
`TCGGCTA; CV-195: GGATC AGAAGGAAAGCAA-
`CAGCAC, AACGMGATGGAACAAGGGMACT.
`
`102
`
`12. S. A. Karl, B. W. Bowen, J. C. Avise, Genetics, in
`press.
`13. Locations and sample sizes for American oysters
`in the present study are as follows: MA, Woods
`Hole, MA (n = 35); SC, Charleston, SC (23); GA,
`Cumberland Island, GA (33); FL1, New Smyma
`Beach, FL (30); FL2, Stuart, FL (29); FL3, Port
`Charlotte, FL (29); FL4, Panacea, FL (39); FL5,
`Carabelle River, FL (18); and LA, Grand Isle, LA
`(41). Living oysters were collected and placed on
`wet ice for transportation to the laboratory. Total
`cell DNA was extracted from mantle-gonad tissue
`by homogenizing in 50 mM tris-HCI (pH 8.0), 100
`mM EDTA, and 100 mM NaCI followed by one
`cycle of phenol, phenol-chloroform, and chloro-
`form extraction. RNA was removed by ribonucle-
`ase A digestion for 3 hours followed by a repeat of
`the organic extractions described above. We pre-
`cipitated DNA by adjusting the aqueous fraction
`to a final concentration of 300 mM sodium acetate
`and 70% ethanol; it was then vacuum-dried and
`resuspended in 1 x TE (10 mM tris-HCI, pH 8.0, 1
`mM EDTA).
`.l of genomic DNA was amplified in
`14. In general, 1
`a 10i-pi reaction volume with 25 pmol of each
`primer and 2.5 units of Taq polymerase, accord-
`ing to manufacturer's instructions (Promega), with
`the addition of MgCI2 and bovine serum albumin
`9g/gi,
`to a final concentration of 2.5 mM and 0.1
`respectively. PCR cycling parameters consisted
`of an initial denaturation at 950C for 2 min, 40
`cycles of denaturation at 940C for 1 min with
`annealing at 55° to 62°C for 1 min and extension at
`720C for 2 min, and a final extension at 720C for 7
`min.
`15. Approximately 10 pL. of the amplified DNA was
`digested directly with 5 units of each restriction
`enzyme in a 20-gi reaction according to the
`manufacturer's directions. Electrophoresis was
`performed in 2.5% agarose gels stained with
`ethidium bromide (180 ng/ml). Restriction pat-
`
`tems were visualized with shortwave ultraviolet
`light. Polymorphisms were indicated by the gain
`or loss of fragments in the restriction profiles.
`16. To eliminate the possibility that our nuclear loci
`mistakenly may have represented mtDNA poly-
`morphisms, we probed a Southem (DNA) blot of
`the amplified scnDNA products with purified oys-
`ter mtDNA. No bands appeared in the autoradio-
`gram except in control lanes. The hypothesis of
`mtDNA contamination is further discounted by the
`nature of the scnDNA polymorphisms themselves,
`which involved diploid genotypes with general
`conformance to Hardy-Weinberg expected geno-
`typic frequencies. This result further supports the
`idea that the loci are inherited in a Mendelian
`fashion. In addition, all pairwise comparisons of
`the four loci show insignificant deviations from
`gametic equilibrium [B. S. Weir and C. C. Cock-
`erham, in Mathematical Evolutionary Theory, M. E.
`Feldman, Ed. (Princeton Univ. Press, Princeton,
`NJ, 1989), pp. 86-110].
`17. D. W. Garton, R. K. Koehn, T. M. Scott, Genetics
`108, 445 (1984).
`18. R. K. Koehn and P. M. Gaffney, Mar. Biol. 82, 1
`(1984).
`19. W. J. Diehl and R. K. Koehn, ibid. 88, 265 (1985).
`20. S. M. Singh and E. Zouros, Evolution 32, 342
`(1978).
`21. J. B. Mitton and R. K. Koehn, J. Exp. Mar. Biol.
`Ecol. 90, 73 (1985).
`22. R. K. Koehn and T. J. Hilbish, Am. Sci. 75, 134
`(1987).
`23. R. C. Lewontin and J. Krakauer, Genetics 74, 175
`(1973).
`24. M. Nei, ibid. 89, 583 (1978).
`25. We thank M. Schexnayder and S. McAlpine for
`oyster specimens. Supported by an NIH predoc-
`toral training grant to S.A.K. and by NSF grant
`BSR-9005940.
`
`4 December 1991; accepted 14 February 1992
`
`Identification of ras Oncogene Mutations in the
`Stool of Patients with Curable Colorectal Tumors
`David Sidransky, Takashi Tokino, Stanley R. Hamilton,
`Kenneth W. Kinzler, Bernard Levin, Philip Frost, Bert Vogelstein*
`Colorectal (CR) tumors are usually curable if detected before metastasis. Because genetic
`alterations are associated with the development of these tumors, mutant genes may be
`found in the stool of individuals with CR neoplasms. The stools of nine patients whose
`tumors contained mutations of K-ras were analyzed. In eight of the nine cases, the ras
`mutations were detectable in DNA purified from the stool. These patients included those
`with benign and malignant neoplasms from proximal and distal colonic epithelium. Thus,
`colorectal tumors can be detected by a noninvasive method based on the molecular
`pathogenesis of the disease.
`
`Colorectal cancer is the third most com-
`world,
`with
`the
`mon malignancy in
`570,000 new cases expected in 1992. In the
`United States alone, over 60,000 people
`
`D. Sidransky, T. Tokino, K. W. Kinzler, B. Vogelstein,
`Department of Oncology, The Johns Hopkins Univer-
`sity, Baltimore, MD 21231.
`S. R. Hamilton, Department of Pathology and Depart-
`ment of Oncology, The Johns Hopkins Hospital, Balti-
`more, MD 21205.
`P. Frost, Department of Cell Biology, M. D. Anderson
`Hospital, Houston, TX 77054.
`B. Levin, Section of Gastrointestinal Oncology, M. D.
`Anderson Hospital, Houston, TX 77030.
`*To whom correspondence should be addressed.
`
`SCIENCE * VOL. 256
`
`*
`
`3 APRIL 1992
`
`will die from colorectal cancer this year (1).
`Whereas individuals with advanced disease
`have a poor prognosis, colorectal tumors
`diagnosed at an earlier stage can usually be
`cured by surgical or colonoscopic excision
`(2). Methods to detect surgically resectable
`tumors could therefore reduce deaths from
`this disease (3). The only noninvasive test
`for such a purpose involves testing stool for
`blood, but the appearance of hemoglobin in
`stool is not specific for neoplasia (4, 5).
`Tumor-derived mutations in oncogenes
`and suppressor genes potentially provide
`more specific markers (6). Mutations in
`
`Downloaded from https://www.science.org on January 10, 2024
`
`Geneoscopy Exhibit 1035, Page 1
`
`

`

`..7t;
`
`these genes appear to be responsible for the
`initiation and progression of most human
`tumors, including those of the colon and
`rectum (7). Theoretically, colorectal tumors
`should shed cells containing these mutations
`into the stool. However, stool is a complex
`mixture consisting of diverse microorganisms,
`undigested food residues, mucus, and soluble
`and insoluble products of the gastrointestinal
`
`Mutant
`probe 2
`
`Wild-type
`
`Patient
`
`Mutant
`probe 1
`
`-
`
`2 1
`
`0 . .
`
`Flg. 1. Identification of ras mutations in stool by
`plaque hybridization. PCR products that con-
`tained the first coding exon of K-ras were
`generated from stools of patients 1, 2, and 10
`(Table 1) and cloned into a bacteriophage
`vector (16, 17). The plaque lifts were hybrid-
`ized to an oligonucleotide specific for wild-type
`Ras, an oligonucleotide specific for Val12 (mu-
`tant probe 1), and to an oligonucleotide specif-
`ic for Asp13 (mutant probe 2), as indicated (20).
`The number of plaques used for hybridization
`to the mutant specific oligonucleotides was
`fivefold higher than that used for the wild-type-
`specific oligonucleotide so that a statistically
`significant number of hybridizing plaques was
`obtained. The ratios of plaques hybridizing to
`the mutant-specific probe compared to the
`wild-type-specific probe were 0.08:1 and
`0.04:1 in patients 1 and 2, respectively.
`
`3
`
`Stool
`4 11
`1
`S
`
`.cPatientb.
`
`11
`
`Tumor
`4
`1
`
`_.
`._
`
`3
`
`vaj12
`
`..
`
`..~A-
`
`WRE
`
`tract. Furthermore, it contains numerous deg-
`most reproducible procedure was used (18).
`radative enzymes derived from cells, food, and
`Approximately 100 mg of stool frozen at
`bacteria. It was therefore unclear whether
`-80°C was diluted with 300 ,l of lysis
`mutant genes from tumor cells could survive
`buffer 1500 mM tris, 16 mM EDTA, 10 mM
`in this hostile environment and be detectable
`NaCI (pH 9.0)], and particulates and most
`in clinical specimens.
`bacteria were removed by centrifugation.
`To investigate this possibility, we exam-
`Proteins were digested with proteinase K
`ined stools from individuals with CR tumors
`and extracted with phenol and chloroform.
`for mutations of K-ras at codons 12 or 13,
`After ethanol precipitation, the DNA was
`which occur commonly in these neoplasms
`further purified by binding to glass beads.
`(8-12). We first analyzed tumors from 24
`From 0.5 to 5.0 ,ug of DNA was typically
`patients for the presence of K-ras gene mu-
`obtained. The first exon of K-ras was then
`tations. These cases comprised randomly
`PCR-amplified from this DNA as described
`chosen individuals from our clinics from
`above. Because we initially expected that
`whom we obtained stool samples before
`mutant ras would represent only a small
`bowel preparation for colonoscopy or surgery
`fraction of the total ras DNA in stool (if
`and who were subsequently found to have
`present at all), we used a sensitive tech-
`either a malignant colorectal tumor (carci-
`nique for analysis. This technique allowed
`noma) or a benign tumor (adenoma) greater
`the identification of a small fraction of
`than 1 cm in diameter. Adenomas of this
`mutant p53 genes in the urine of patients
`with advanced bladder cancers and can
`size are clinically the most important, as
`they are much more likely to progress to
`reveal the existence of even one mutant
`malignancy than smaller tumors (13, 14).
`gene among several thousand normal genes
`The first exon of K-ras gene was ampli-
`(19). The PCR products were cloned in a
`fied from DNA purified from cryostat sec-
`bacteriophage vector and the phage DNA
`tions of these tumors (15) by the polymer-
`transferred to nylon filters (20). These fil-
`ase chain reaction (PCR), (16). The PCR
`ters were then incubated with 32P-labeled
`products were cloned, and pooled clones
`oligonucleotides
`that
`recognized
`either
`were sequenced to identify mutations (17).
`wild-type K-ras, the mutant K-ras found in
`Nine of the 24 tumors (37%) were found to
`the tumor, or another mutant K-ras gene as
`contain mutations of this exon. Three dif-
`a negative control. With this assay, we
`ferent mutations were identified (Gly'2 to
`found that both patients contained mutant
`Val'2 or Asp'2; and Gly'3 to Asp'3). These
`ras in the DNA purified from their stool
`data were consistent with previous studies
`samples. The mutant genes detected in the
`that showed ras gene mutations in about
`stool were the same as those detected in the
`50% of such tumors, with 84% of the
`tumors (Val'2 in the stool and tumor of
`mutations confined to codons 12 or 13 of
`patient 1; Asp'3 in the stool of patient 2,
`K-ras (10).
`1). A control stool sample from a
`Fig.
`We next analyzed the stools from the
`patient without a ras mutation in his tumor
`first two of the nine patients. Several meth-
`contained no mutation at either of these
`ods to purify DNA were evaluated, and the
`positions (Fig. 1, patient 10).
`Table 1. Patients studied for stool ras gene mutations.
`
`Mutant
`Tumnor
`muta-
`ras
`tmiota-
`inge
`tin nstoolt
`+
`+
`+
`+
`
`Pa-
`tient
`
`Age/
`sex
`
`Tumor
`location
`
`Tumor
`type/stage*
`
`Tumor size
`(cm3)
`
`5.8 x 6.5 x 2.7
`1.5 x 1.5 x 0.6
`2.8 x 2.0 x 0.4
`2.5 x 4.7 x 1.8
`1.0 x 0.9 x 0.4
`5.9 x 6.4 x 1.7
`4.3 x 3.4 x 1.4
`4.8 x 3.0 x 1.2
`6.0 x 4.011
`12 x 6.511
`7 x 3 x 2
`2.1 x 3.1 x 0.4
`
`Val12
`Carcinoma/C
`52/F
`Rectum
`1
`Asp13
`Sigmoid colon
`63/M
`Adenoma
`2
`Asp12
`Carcinoma/C
`51/F
`Rectum
`3
`Asp12
`Carcinoma/C
`61/M
`Rectum
`4
`Asp13
`Carcinoma/A
`70/M
`Rectum
`5
`Asp12
`Carcinoma/B
`71/M
`Rectum
`6
`+
`Asp12
`Carcinoma/B
`51/F
`Ascend colon
`7
`+
`Asp13
`69/M
`Sigmoid colon
`Carcinoma/B
`8
`+
`Asp12
`Cecum
`67/M
`Adenoma
`9
`+
`Carcinoma/C
`Splenic flex
`None
`61/M
`10
`-
`Carcinoma/B
`Sigmoid colon
`None
`34/M
`11
`-
`Carcinoma/C
`Sigmoid colon
`63/M
`None
`12
`-
`42/F
`NA§
`NA
`NA
`13
`-
`NA
`NA§
`53/F
`NA
`14
`-
`NA
`NA§
`63/F
`NA
`15
`-
`*Carcinomas were classified according to Duke (13): A, confined to muscularis propria; B, extension through
`tDetermined by sequence
`muscularis propria, but confined to colon; C, metastatic to regional lymph nodes.
`tAs assessed by the plaque hybridization or Southern blot (DNA) assay
`analysis of codons 12 or 13 of K-ras.
`with three mutant specific oligomers (Val12, Asp12, and Asp13) as probes.
`§Patients 13, 14, and 15 had no
`1lOnly measurements of the external
`colorectal neoplasms found at colonoscopy (NA = not applicable).
`surfaces of these tumors were available.
`
`SCIENCE * VOL. 256
`
`*
`
`3 APRIL 1992
`
`103
`
`Fig. 2. Southern blot assay for ras mutations.
`PCR products that contained the first coding
`exon of K-ras were generated from stools and
`tumors of patients 1, 3, 4, and 1 1, as indicated
`(16). PCR products were subjected to electro-
`phoresis through a 2% agarose gel and trans-
`(21). The blots were
`ferred to nylon filters
`hybridized to oligonucleotide probes specific
`for wild-type ras (bottom), the Asp12 mutation
`(middle), or the Val12 mutation (top).
`
`Downloaded from https://www.science.org on January 10, 2024
`
`Geneoscopy Exhibit 1035, Page 2
`
`

`

`~.
`
`lit
`
`"' gm
`go
`
`The fraction of phage plaques hybridiz-
`ing to the mutant-specific oligonucleotide
`in patients 1 and 2 was high, representing 8
`and 4%, respectively, of the phage hybrid-
`izing to the oligonucleotide specific for the
`wild-type K-ras. This suggested that a less
`sensitive, but simpler, assay could be used
`to identify mutant genes in stool samples.
`For this purpose, crude PCR products were
`simply subjected to electrophoresis through
`an agarose gel and transferred to nylon
`filters by the method of Southern (21).
`These blots were then incubated with 32p_
`labeled oligonucleotides
`that recognized
`wild-type or mutant K-ras. The Val12 mu-
`tation of patient 1 was easily observed in
`stool with the Southern blot (DNA) assay
`(Fig. 2, top). The oligonucleotide specific
`for the Asp12 mutant provided a negative
`control (Fig. 2, middle). The wild-type-
`specific oligonucleotide hybridized to DNA
`from both tumor and stool, as expected
`(Fig. 2, bottom). Similarly, Southern blot
`analysis revealed that the tumors and stool
`of the DNA from patients 3 and 4 both
`contained the Asp12 mutation, whereas
`neither hybridized to the Val12-specific oli-
`gonucleotide (Fig. 2). The ratio of mutant
`to wild-type hybridization in the stool sam-
`ples was five- to tenfold lower than that in
`the tumors, consistent with the plaque hy-
`bridization assays.
`Southern blot analysis detected muta-
`tions in stool from eight of the nine patients
`(Table 1). Mutations originating in benign
`tumors (patients 2 and 9) and malignant
`tumors were detected. Tumors as small as
`1.3 cm3 gave rise to detectable mutant
`genes in the stool (patient 2). Moreover,
`proximal as well as distal tumors (patient 9,
`cecum; patient 7, ascending colon) yielded
`positive results.
`As controls, we examined six stool sam-
`ples, three from individuals with no colo-
`rectal neoplasia and three from individuals
`with colorectal tumors that did not contain
`K-ras mutations at codons 12 or 13. In all
`six cases, strong hybridization to the wild-
`type-specific oligonucleotides, but not to
`specific
`for Val'2,
`the oligonucleotides
`Asp12, or Asp13 mutations, was observed
`(Table 1 and Figs. 1 and 2).
`We were surprised at the ease with
`which K-ras mutations were identified in
`the stool. However, rough calculations in-
`dicate that colorectal tumors could consti-
`tute a significant fraction of the human
`DNA present in stool. The colon is more
`than a meter in length, but the epithelium
`is confined to a lining only a few millime-
`ters thick. It is estimated that the normal
`adult colon contains 5 x 1010 epithelial
`cells (22). One-sixth to one-third of these
`are shed every 24 hours, giving rise to
`approximately 1010 normal cells per day
`(23). A tumor of 1 cm3 may contain more
`104
`
`than 109 cells. These cells turn over at
`similar or elevated rates compared to nor-
`mal cells (24). Thus, it is conceivable that
`more than 1% of the epithelial cells shed
`from colon into the stool could be derived
`from tumors of this size. It is also possible
`that tumor cells are more resistant to the
`degradative processes in feces or that apop-
`tosis, which degrades DNA in normally
`differentiating cells, is not fully operative in
`tumor cells.
`Mutations of ras are particularly amena-
`ble for these studies because they are pre-
`sent in both benign and malignant CR
`tumors and mutations occur at relatively
`few codons (S-12). However, ras mutations
`are found in only 50% of such tumors.
`Other mutant genes present in colorectal
`tumors (7) could probably be detected in
`the stool, which would increase the poten-
`tial sensitivity of this strategy. With the
`help of ras and other gene probes, addition-
`al studies, in which the stools from a large
`number of patients with colorectal tumors
`of varying size, stage, and anatomical posi-
`tion are analyzed, will be needed to more
`accurately determine the sensitivity of such
`tests.
`These results provide the conceptual
`and practical basis for a new approach for
`detecting the presence of colorectal tumors
`in a noninvasive fashion. The approach
`could have use in monitoring patient pop-
`ulations on different diets or treatments
`designed to minimize the incidence of co-
`lorectal neoplasia (25). It also could even-
`tually find use in screening asymptomatic
`patients, especially those at risk by virtue of
`inherited or environmental factors
`(25,
`26), for the presence of colorectal neopla-
`sia. An analysis of data from only ras probes
`suggests that some early colorectal cancers
`and premalignant lesions might be identifi-
`able through this strategy. Because colorec-
`tal tumors are so common, this approach
`has implications for public health.
`
`REFERENCES AND NOTES
`
`1. M. Shike et al., Bull. WHO 68, 377 (1990); R. W.
`Beart, Jr., in Textbook of Clinical Oncology, A. I.
`Hollieb, D. S. Fink, G. P. Murphy, Eds. (American
`Cancer Society, Atlanta, 1991), pp. 213-218.
`2. More than 90% of patients with adenomas or with
`carcinomas confined to the mucosa can be cured
`by surgical methods. About 40 to 80% of patients
`whose disease is confined to local extension or
`lymph node involvement are similarly curable, but
`less than 10% of patients with distant metastasis
`can be treated successfully. T. Rich, L. L. Gun-
`derson, J. Galdabin et al., Cancer 52, 1317
`(1983); C. E. Dukes and H. J. R. Bussey, Br. J.
`Cancer 12, 309 (1958); E. M. Copeland, L. D.
`Miller, R. S. Jones, Am. J. Surg. 116, 875 (1968);
`R. S. Grinnell, Ann. Surg. 2, 494 (1950); B. Eisen-
`berg et al., Cancer 49, 1131 (1982); D. A. August,
`R. T. Ottow, P. H. Sugarbaker, Cancer Metastasis.
`Rev. 3, 303 (1984).
`3. S. J. Winawer, D. Schottenfield, B. S. Frehinger, J.
`Natl. Cancer Inst. 83, 243 (1991).
`4. D. F. Ransohoff and C. A. Lang, N. Engi. J. Med.
`
`SCIENCE * VOL. 256
`
`*
`
`3 APRIL 1992
`
`325,37 (1991); J. J. DeCosse and J. S. Jacobson,
`ibid. 326,134 (1991); J. N. Cooper, K. O'Connor,
`S. J. Winawer, ibid., p. 136.
`5. D. M. Eddy et al., Gastroenterology 92, 682
`(1987); D. M. Eddy, Ann. Intern. Med. 113, 373
`(1990); A. B. Miller et al., Int. J. Cancer 46, 761
`(1990).
`6. J. M. Bishop, Cell 64, 235 (1991); R. A. Weinberg,
`Science 254, 1138 (1991).
`7. E. R. Fearon and B. Vogelstein, Cell 61, 759
`(1 990).
`8. J. L. Bos et al., Nature 327, 293 (1987).
`9. K. Forrester, C. Almoguera, K. Han, W. E. Grizzle,
`M. Perucho, ibid., p. 298.
`10. B. Vogelstein et al., N. Engi. J. Med. 319, 525
`(1988).
`11. C. J. Farr etal., Oncogene3, 673 (1988).
`12. P. Delattre et al., Lancet 335, 353 (1989).
`13. A. M. Cohen, B. Shank, M. A. Friedman, in Can-
`cer: Principles and Practice of Oncology, V. T.
`DeVita, S. Hellman, S. A. Rosenberg, Eds. (Up-
`pincott, Philadelphia, ed. 3,1989), pp. 895-964.
`14. S. Stryker et al., Gastroentology 93, 1009 (1987);
`T. Muto, H. J. R. Bussey, B. C. Morson, Cancer
`36, 2251 (1975); B. C. Morson, ibid. 34, 845
`(1974); H. E. Knoernschild, Surg. Forum 14, 137
`(1963); S. Welin et al., Am. J. Radiol. 90, 673
`(1963); D. F. Ransohoff and C. A. Lang, J. Am.
`Med. Assoc. 264, 76 (1990).
`15. E. R. Fearon, A. P. Feinberg, S. R. Hamilton, B.
`Vogelstein, Nature 318, 377 (1985).
`16. The sense primer was 5'-AGGAATTCATGACT-
`GAATATAAACTTGT-3', and the antisense primer
`was 5'-ATCGAATTCCTCTATTGTTGGATCATAT-
`TC-3'. These primers included Eco RI sites at their
`5' ends to facilitate cloning. PCR was performed
`[as described in S. J. Baker et al., Cancer Res. 50,
`7717 (1990)]. Thirty-five cycles were performed
`for tumor DNA and 45 cycles for stool DNA.
`17. PCR products were purified by phenol-chloroform
`extraction and ethanol precipitation. They were
`cleaved with Eco RI and repurified, and approxi-
`mately 50 ng of DNA was ligated to lambda Zap II
`vector arms (Stratagene) and packaged accord-
`ing to the manufacturer's directions. XL1 Blue
`cells (Stratagene) were infected with bacterio-
`phage, and double-stranded plasmids were ob-
`tained using a helper phage as in J. M. Nigro et al.
`[Nature 342, 705 (1989)]. At least 100 clones were
`pooled for sequencing with the primer 5-AT-
`TCGTCCACAAAATGAT-3'.
`18. The stool-lysis buffer mix was vortexed for 30 s
`and cleared by centrifugation at 12,000g for 2
`min. DNA in the supernatant was purified by
`SDS-proteinase K digestion, phenol-chloroform
`extraction, and ethanol precipitation
`[as de-
`scribed in S. E. Goeiz, S. R. Hamilton, B. Vo-
`gelstein, Biochem. Biophys. Res. Commun. 130,
`1 18 (1985)]. The DNA from stool was then further
`purified by binding to glass [B. Vogelstein and D.
`Gillespie, Proc. Nati. Acad. Sci. U.S.A. 76, 615
`(1978)].
`19. D. Sidransky et al., Science 252, 706 (1991).
`20. XL1 cells infected with bacteriophage containing
`PCR products were plated on L-agar at a density
`of 100 to 2000 plaques per plate, transferred to
`nylon membranes, and hybridized with oligonu-
`cleotides specific for wild-type or mutant K-ras.
`The oligonucleotides used for hybridization were
`5'-GGAGCTGGTGGCGTAGGCAA-3' for
`wild-
`type ras, 5'-GGAGCTGTTGGCGTAGGCAA-3' for
`the Val12 mutant, 5'-GGAGCTGATGGCGTAG-
`GCAA-3' for the Asp12 mutant, and 5'-GGAGC-
`TGGTGACGTAGGCAA-3' for the Asp13 mutant.
`Oligonucleotides were labeled with 32p to a spe-
`cific activity of approximately 1 o8 dpm/Lg with T4
`polynucleotide kinase and hybridized as in (19).
`21. After electrophoresis through a 2% agarose gel,
`the gel was incubated in 16 mM HCI for 30 min
`and transferred to Zetabind filters (Bio-Rad) in 0.4
`M NaOH [as in K. C. Reed and D. A. Mann,
`Nucleic Acids Res. 13, 7207 (1985)]. Filters were
`hybridized as in (19).
`22. L. E. Mehi, J. Surg. Oncol. 47, 243 (1991).
`23. R. G. Shorter et al., Am. J. Digest. Dis. 9, 760
`
`Downloaded from https://www.science.org on January 10, 2024
`
`Geneoscopy Exhibit 1035, Page 3
`
`

`

`----------------------
`
`(1964); M. Lipkin, B. Bell, P. Shercock, J. Clin. Inv.
`42, 767 (1963); W. C. MacDonald, J. S. Trier, N. B.
`Everett, Gastroenterology 46, 405 (1964).
`24. B. Tribukait, C. Hammerberg, C. Rubio, Acta
`Pathol. Microbiol. Immunol. Scan. Sect. A 91, 89
`(1983); P. Quirke etal., J. Pathol. 151, 285 (1987);
`C. Hammarberg, B. Tribukait, U. Ohman, Acta
`Radiol. Oncol. 25, 45 (1986).
`25. M. Lipkin, Cancer Res. 48, 235 (1988); H. L.
`Newmark, M. J. Wargovich, W. R. Bruce, J. Nati.
`Cancer Inst. 72, 1323 (1984).
`
`26.
`
`27.
`
`J. Utsunomiya and H. T. Lynch, Eds., Hereditary
`Colorectal Cancer (Springer-Verlag, Tokyo, 1990).
`We thank J. Sitzmann, R. Fishbein, and K. D.
`Lillemoe for providing clinical samples and P.
`Green, C. R. Robinson, A. C. Preisinger, B. Ender,
`and M. C. Olive for technical assistance. Support-
`ed by the Clayton Fund, the McAshan Fund, and
`grants CA06973 and CA35494 from the National
`Cancer Institute of NIH.
`
`17 January 1992; accepted 2 March 1992
`
`CD1 9: Lowering the Threshold for Antigen
`Receptor Stimulation of B Lymphocytes
`Robert H. Carter* and Douglas T. Fearon
`Lymphocytes must proliferate and differentiate in response to low concentrations of a vast
`array of antigens. The requirements of broad specificity and sensitivity conflict because the
`former is met by low-affinity antigen receptors, which precludes achieving the latter with
`high-affinity receptors. Coligation of the membrane protein CD1 9 with the antigen receptor
`of B lymphocytes decreased the threshold for antigen receptor-dependent stimulation by
`two orders of magnitude. B lymphocytes proliferated when approximately 100 antigen
`receptors per cell, 0.03 percent of the total, were coligated with CD1 9. The B cell resolves
`its dilemma by having an accessory protein that enables activation when few antigen
`receptors are occupied.
`
`The immune system must respond to low
`concentrations of antigen for the efficient
`elimination of infections. The growth and
`differentiation of lymphocytes are mediated
`by antigen receptors that have low affinity
`for their ligands because they are products
`of recombinatorial gene rearrangement that
`takes place in the absence of selection by
`antigen. Therefore, lymphocytes must have
`mechanisms that enable them to be stimu-
`lated when relatively few receptors have
`bound antigen. The T lymphocyte has the
`accessory membrane proteins CD4 and
`CD8, which, when coligated with the T
`cell antigen receptor (TCR) by the major
`histocompatibility complex (MHC)-pep-
`tide complex, decrease the number ofTCRs
`that must be ligated (1). No membrane
`protein of the B lymphocyte that is analo-
`gous to CD4 and CD8 has been identified.
`The CD19 membrane protein, a mem-
`ber of the immunoglobulin (Ig) superfam-
`ily, is B cell-specific and is expressed at
`each developmental stage except that of the
`terminally differentiated plasma cell (2). It
`is a component of a complex that contains
`at least two other membrane proteins, com-
`plement receptor type 2 (CR2, also called
`CD21) and TAPA-1 (3). The CR2 protein
`mediates the capacity of the complement
`system to enhance the production of anti-
`body in response to low concentrations of
`Division of Molecular and Clinical Rheumatology, De-
`partment of Medicine, and the Graduate Program in
`Immunology, Johns Hopkins University School of
`Medicine, Baltimore, MD 21205.
`*To whom correspondence should be addressed.
`
`antigen in vivo (4). Both CD19 and CR2
`augment activation of phospholipase C in B
`cells when coligated with membrane Ig
`(mIg), but independent ligation of CD19
`suppresses B cell activation (5, 6).
`To determine whether ligating CDl9 al-
`ters the capacity of mIgM to induce DNA
`synthesis, replicate samples of human B cells
`and mitomycin-treated murine L cells that
`express human FcYRII (FczRII-L cells) (7)
`were cultured for 2.5 days in the presence of
`incremental concentrations of a monoclonal
`antibody (MAb) to IgM, a saturating con-
`centration of MAb to CD19, or an equal
`irrelevant MAb,
`of an
`concentration
`MOPC-21, and interleukin-4 (IL-4). In the
`absence of anti-CD19, the lowest concen-
`tration of anti-IgM that induced incorpo-
`
`.50,
`
`CL
`
`g40O
`
`30,
`
`20
`
`C
`
`/
`
`ni1
`
`10-14
`
`1012
`
`1T10
`AtMws (M)
`
`10-8
`
`10-6
`
`ration of I3Hlthymidine by B cells above the
`background incorporation was 6.7 x 10`1
`M. In the presence of anti-CD19, this was
`reducedto6.7 x 10-3M (Fig. 1) (8, 9).At
`higher concentrations of anti-IgM, anti-
`CD19 also increased the magnitude of
`[3H~thymidine incorporation to twice that
`induced by anti-IgM alone. Anti-CD19
`alone did not induce proliferation. The liga-
`tion of CD19 both lowered the threshold for
`B cell activation by mIgM and increased the
`magnitude of the B cell response to optimal
`levels of mIgM stimulation.
`In a parallel experiment to assess the
`binding characteristics of the monoclonal
`anti-IgM, replicate samples of B cells and
`FcYRII-L cells were incubated with incre-
`mental concentrations of `251-labeled anti-
`IgM in the presence or absence of 2 x 10-7
`M unlabeled anti-IgM. Cell-bound and free
`antibody were separated, and specific bind-
`ing was calculated. The L cells alone did not
`specifically bind the radiolabeled antibody.
`Scatchard analysis showed that the anti-IgM
`bound to the mixture of B cells and L cells
`with a dissociation constant (Kd) of 1.8 X
`10-9 M, and to 2.7 x 105 sites per cell at
`saturation. The presence of anti-CD19 did
`not alter the Kd of the anti-IgM for B cells in
`the presence of FcYRII-L cells (10). There-
`fore, in Fig. 1, the threshold concentration
`of anti-IgM that induced B cell proliferation
`in the presence of anti-CD19 bound to only
`92 mIgM molecules per cell, or 0.03% of the
`total mIgM expressed per cell.
`We measured the fraction of CD19 that
`must be ligated to augment B cell prolifer-
`ation by incubating replicate samples of B
`cells and FcYRII-L cells in the presence of
`suboptimal anti-mIgM, IL-4, and incre-
`mental concentrations of anti-CD19 for 2.5
`days, after which incorporation of I3Hlthy-
`midine was assayed. The enhancing effect
`of anti-CD19 increased incrementally from
`6.7 x 10`2 M, which bound <0.2% of
`CD19, up to a saturating concentration of
`6.7 x 10-8 M (Fig. 2A). To determine
`
`Fig. 1. Enhancement of DNA synthesis in B
`lymphocytes by the coaggregation of migM
`and CD19. Replicate samples of 5 x 104 pe-
`ripheral blood B cells (8) were cultured in 0.2
`ml of RPMI-1640 with 10% fetal calf serum in
`flat-bottom culture plates with 2 x 104 mitomy-
`cin-treated fibroblastic L cells expressing hu-
`man CDw32/FczRI and recombinant IL-4 (200
`units/mI, Genzyme). The B cells were stimulat-
`ed with a range of concentrations of MAb
`DA4.4 anti-lgM (9) together with either control
`antibody MOPC-21 (0) or HD37 anti-CD19 (-).
`Cells were pulsed with [3H]thymidine for the
`last 16 hours of a 60-hour culture. Results are
`means ± SD of triplicates. Cells cultured with
`either control antibody or anti-CD19 alone in-
`corporated 2726 ± 256 and 2475 ± 255 counts
`per minute (cpm), respectively. Representative
`of four experiments.
`
`SCIENCE * VOL. 256
`
`*
`
`3 APRIL 1992
`
`105
`
`Downloaded from https://www.science.org on January 10, 2024
`
`Geneoscopy Exhibit 1035, Page 4
`
`

`

`Identification of ras Oncogene Mutations in the Stool of Patients with Curable
`Colorectal Tumors
`David Sidransky, Takashi Tokino, Stanley R. Hamilton, Kenneth W. Kinzler, Bernard Levin, Philip Frost, and Bert Vogelstein
`
`Science 256 (5053), . DOI: 10.1126/science.1566048
`
`View the article online
`https://www.science.org/doi/10.1126/science.1566048
`Permissions
`https://www.science.org/help/reprints-and-permissions
`
`Use of this article is subject to the Terms of service
`
`Science (ISSN 1095-9203) is published by the American Association for the Advancement of Science. 1200 New York Avenue NW,
`Washington, DC 20005. The title Science is a registered trademark of AAAS.
`
`Downloaded from https://www.sci